AI ASSISTED

LITERATURE
REVIEW
WRITING

Interactive eBook with insightful links, engaging

videos and lifetime content updates

Faharol Zubir
 AI Assisted
Literature Review
Writing

Faharol Zubir
AI Assisted Literature Review Writing

Author: Faharol Zubir

Editor: Norlizawati Ghazali
Layout designer: Suraya Najieha Kamarudin
Book cover design: Syafiq Fizieri Kamarudin

Copyright © 2024 by Faharol Zubir. All rights reserved.

No part of this publication may be reproduced, distributed, or

transmitted in any form or by any means, including photocopying,
recording, or other electronic or mechanical methods, without the
prior written permission of the publisher, except in the case of brief
quotations embodied in critical reviews and certain other non-
commercial uses permitted by copyright law.

For permission requests, write to the publisher, addressed

“Attention: Permissions Coordinator,” at the address below:

[email protected]
faharol zubir

First Edition: January 2024

ISBN:
Preface
In academic research, the integration of artificial
intelligence (AI) into the systematic process of
literature review marks a significant transformation.
As scholars, researchers, and students grapple with
the increasing expanse of published knowledge, AI
emerges not as an usurper of human intellect, but as
a collaborator extending the reach of human capacity.
This book is conceived as a bridge between the
traditional diligence of academic inquiry and the
cutting-edge assistive potential of AI. It is designed to
be a guide, a companion, and a resource that
illuminates the path of literature review, a
foundational aspect of research that demands both
breadth and depth of understanding.

The craft of literature review is a rigorous one,

involving the identification, synthesis, and analysis of
relevant scholarly material. It is an exercise in critical
thinking, precision, and extensive reading. Yet, it is
no longer a process that must be undertaken alone.
AI has the potential to be a valuable ally in this
endeavor, amplifying the researcher's abilities to
search, manage, comprehead and write. Through the
chapters of this book, we will explore how AI can aid
each of these stages without compromising the
essential human touch that is at the heart of all
scholarly work.

We begin with the fundamentals of literature search,

reviewing into the conventional methods before
exploring how AI can enhance this process. We look
at the development of search strings and the
sophisticated ways AI can assist in their creation,
offering examples and strategies for effective
literature retrieval. In discussing literature
management, we examine how AI can bring
organization and clarity to the overwhelming task of
cataloging and synthesizing research findings.

Comprehension is the next frontier, where we

consider the role of AI in aiding the understanding of
complex texts, leveraging tools that can provide
insights and summaries to support the researcher's
engagement with the material. In the final section, we
address the culmination of the literature review
process that is writing. Here, we focus on how AI can
assist in crafting drafts and outlines, providing
support that enhances the writer's own analytical and
narrative skills.
This book is not only a guide but also a reflection of a
profound belief: that AI should be an assistant, not a
replacement, in the intellectual pursuit of knowledge.
It is my hope that readers will find within these pages
not only practical guidance but also a vision for a
collaborative future where AI and human intelligence
work in concert to further the frontiers of knowledge.

As you navigate through the chapters, enriched with

tutorial videos for a more comprehensive
understanding, I invite you to approach AI as you
would a trusted colleague: with openness to its
capabilities and with the discernment to integrate its
assistance into your scholarly work judiciously. Let
this book be a testament to the harmonious potential
of human and artificial intelligence, working together
to achieve the timeless goal of knowledge creation
and dissemination.

Welcome to a journey that respects tradition while

embracing innovation. A journey through the art and
science of AI-assisted literature review.

How to use this eBook.

TABLE OF
CONTENTS

1 INTRODUCTION
Introduction 1

2 PURPOSE OF CONDUCTING A LITERATURE

6
REVIEW
Understanding the Current State of Research 7
Establishing Context and Relevance 9
Identifying Gaps and Questions 12
Avoiding Duplication 15
Building a Theoretical Framework 16
Methodological Insights 19
Developing Critical Thinking 21
Establishing Academic Credibility 24

Part I: Literature Search 27

3 FUNDAMENTALS OF LITERATURE SEARCH

The Conventional Method of Doing
28
29
Literature Search
Understanding Online Databases 30
Online Databases for Thesis and 30
Dissertations
Research Article Databases for Literature 32
Search
Scopus 33
WOS 35
Google Scholar 36
The Importance of Keywords 38
 1

DEVELOPING AN EFFECTIVE SEARCH

4 STRING 41
Examples of search strings 43
Scopus 43
WOS 46
Google Scholar 47

5 AI AIDED SEARCH STRING DEVELOPMENT 49

Search String Development Using ChatGPT 51
Search String Development Using Google Bard 53
Search String Development Using Perplexity 54
Search String Development Using Claude 54
Literature Search using ChatGPT 56
Literature search using Bard 59
Literature search using Perplexity 65
Literature search using Claude 66

6 THE CONVENTIONAL AND AI ASSISTED

SNOWBALLING STRATEGY 72
Manual Snowballing 73
AI-Assisted Snowballing Strategy 75
Advantages of AI-Assisted Snowballing 75
AI Tools for Snowballing 77
Connected Papers 78
Litmaps 83
Semantic Scholar 89
 1

Research Rabbit 89
Elicit 99
Scite 103
Mendeley 103
Iris.ai 108
Conclusion 109

7 RETRIEVING LITERATURE 112

113
Accessing Academic Literature
AI in Retrieving Literature 114

Part II Literature Management 116

8 ORGANIZING LITERATURE WITH

MENDELEY 117
Step-by-Step Process of Literature 118
Management with Mendeley

9 DEVELOPING THE LITERATURE MATRIX 129

Steps of Developing A Literature Matrix 131

10 AI ASSISTED LITERATURE MATRIX 136

AI Aided Literature Matrix Using ChatGPT 137
The Role of AI in Managing and Organizing 145
Literature
 1

Detailed Exploration of AI-based Literature 145

Management Tools
Strategies for Efficient Literature 146
Categorization Using AI

Part III: Literature Comprehension 147

11 READING SKILLS FOR

LITERATURE REVIEW 148
Literature Review Reading Techniques 149

AI ASSISTED LITERATURE
12 COMPREHENSION 152
Using ChatGPT for Literature Reading and 154
Comprehension
Literature Comprehension Using ChatGPT 3.5 157
Literature Comprehension Using ChatGPT 4 174
(Plugins)
Literature Comprehension Using ChatGPT 178
(Advanced Data Analysis)
Literature Comprehension Using ChatGPT 4 184
(Google Chrome Extension)
Literature Comprehension Using ChatPDF 190
 1

Part IV Literature Review Writing 203

13 WRITING SKILLS FOR LITERATURE

REVIEW 204

14 AI ASSISTED LITERATURE REVIEW

DRAFT 208
Literature Outline Drafts Using ChatGPT 211
Literature Review Drafts Using Scite 231
Literature Review Drafts Using Scite (Assistant) 232
Literature Review Drafts Using Scite (Custom 245
Dashboard)

15 DEVELOPING CUSTOM GPT FOR

LITERATURE REVIEW TASKS 257
What is Custom GPT? 258
Developing Custom GPT 259

16 HARNESSING LITERATURE REVIEW RELATED

GPTS FROM GPT STORE
291

17 CONCLUSION 301
 Table of Content for Video Tutorials
NUM VIDEO TUTORIALS LENGTH

1. eBook introduction 2:09 min

2. How to use this eBook 1:50 min

Part I : Literature Search

3. The importance of keywords 4:24 min

4. Scopus search features 6:51 min

5. Scopus search features (2) 2:16 min

6. WOS search features 3:41 min

AI Assisted Search String

7. 5:11 min
Development ChatGPT

Search string using Bard,

8. 5:33 min
Perplexity and Claude

9. Activating Plugins Features 3:08 min

Literature search using

10. 3:42 min
ChatGPT
NUM VIDEO TUTORIALS LENGTH

11. Literature search using Bard 2:23 min

Literature search using

12. 3:22 min
Perplexity

13. Literature search using Claude 3:01 min

14. Literature search using Scite 3:58 min

15. Manual snowballing 2:46 min

Snowballing for literatures

16. 6:04 min
using connected papers

Snowballing for literatures

17. 4:58 min
using Litmaps

Snowballing for literatures

18. 5:29 min
using Research Rabbit
NUM VIDEO TUTORIALS LENGTH

Snowballing for literatures

19. 4:26 min
using Elicit

Snowballing for literatures

20. using Scite 6:08 min

Part II : Literature
Management

Importing documents to
21. 2:09 min
Mendeley

Identifying metadata
22. 1:17 min
information in Mendeley

23. Creating folders in Mendeley 2:09 min

Reading, highlighting and

24. 1:56 min
inserting notes in Mendeley

25. Citing while you write 3:33 min

NUM VIDEO TUTORIALS LENGTH

Generating bibliographies using

26. 1:02 min
Mendeley

Conventional literature matrix

27. 6:19 min
table

Developing a literature matrix

28. 7:28 min
using ChatGPT

Developing a literature matrix

29. 5:02 min
using ChatGPT 2

Part III: Literature

Comprehension

Literature comprehension using

30. 6:20 min
ChatGPT 3.5

Literature comprehension using

31. 4:30 min
ChatGPT 4 (Plugins)

Literature comprehension using

32. ChatGPT 4 (Advance Data 5:42 min
Analysis)
NUM VIDEO TUTORIALS LENGTH

Literature comprehension using

33. ChatGPT (Google Chrome 6:34 min
Extension)

Literature comprehension using

34. 7:55 min
ChatPDF

Literature comprehension using

35. 6:43 min
ChatPDF (multiple documents)

Conclusion for literature

36. 2:56 min
comprehension

Part IV: Literature Review

Writing

Literature review outline drafts

37. 9:05 min
using ChatGPT

Writing literature review drafts

38. 4:09 min
using ChatGPT

39. Introduction to Scite 3:37 min

NUM VIDEO TUTORIALS LENGTH

40. Scite’s Assistant features 10:18 min

Writing literature review drafts

41. 10:11 min
using Scite (Custom Dashboard)

Creating literature review

42. 15.06 min
related Custom GPT

Harnessing Custom GPT from

43. 9:08 min
GPT Store

44. eBook Conclusion 2:18 min

3 hrs
Total hours for tutorial videos
37 min
I N T R O D U C T I O N
Literature review writing can be compared with the art
of cooking, where the literature search process is
comparable to the careful selection of ingredients
required to create a delectable dish. Just as a chef
sources the best and most relevant ingredients to
ensure the success of a recipe, a researcher
conducts a literature search to gather the most
pertinent and insightful sources that form the
foundation of a literature review.

Finding the Ingredients: The Literature Search

In cooking, the quest for the perfect ingredients

involves visiting different markets, stores, or gardens
to find items that will contribute to the flavour, texture,
and nutritional value of the dish. Similarly, in a
literature search, the researcher explores various
databases, libraries, and repositories to find
publications that will add depth, context, and
substance to the literature review.

Just as a chef knows that the quality of the

ingredients can make or break a meal, a researcher
understands that the relevance and credibility of the
literature sourced are crucial to the integrity and value
of the review. The chef must be selective, selecting

2
only the freshest produce or the most appropriate
herbs and spices. In parallel, the researcher must be
judicious, choosing only the most pertinent and
academically rigorous texts that align with their
research questions and objectives.

Preparing the Ingredients: Analysing the

Literature

Once the ingredients are sourced, a chef will carefully

prepare them, considering how each will contribute to
the final dish. They chop, measure, and season, all
while keeping in mind the harmony of flavours. In a
literature review, this is the same with analysing and
synthesizing the information gathered. The
researcher evaluates each source, determining its
contribution to the field, and meticulously organizes
the information to present a coherent narrative, much
like a chef assembles ingredients to present a
harmonious and flavourful dish.

Cooking the Dish: Writing the Literature Review

The act of cooking combines ingredients under the

right conditions to create something new and
delightful. Similarly, writing the literature review is

3
where the 'cooking' happens in academia. The
researcher combines all the 'ingredients' such as the
various studies, theories, and findings in order to
create a 'dish' that is rich in information and analysis.
The literature review should be well-balanced, with
each 'ingredient' or source adding something
essential to the 'flavour' or argument of the piece.

Tasting and Refining: Revising the Literature

Review

In cooking, taste-testing is essential. The chef

samples the dish, adjusting ingredients to perfect the
taste. In literature review writing, this is the revision
process. The researcher reviews their work, ensuring
that each part of the 'meal' works together, that the
'flavours' are balanced, and that the 'dish' as a whole
is satisfying and complete. They may find that
additional 'ingredients' are needed, perhaps a study
they missed or a theory that would add depth to their
analysis.

4
Presentation: Finalizing the Literature Review

Finally, the chef plates their dish, considering the

presentation's visual appeal because we eat with our
eyes first. The researcher similarly ensures that the
literature review is well-structured and presented in a
format that is appealing and accessible to the reader,
inviting them to enjoy the scholarly feast laid out
before them.

Before we go further into details of conducting the

processes involved in literature review such as
literature search, management, comprehension and
writing, I believe it is best to revise on the purpose of
conducting a literature review. This is so that when
we go into the processes in detail, we will have a
clearer understanding of the strategies that can be
employed in conducting a literature review.

The video of the introduction for the eBook can be

accessed using the link below

eBook introduction

5
 P U R P O S E O F
C O N D U C T I N G
A L I T E R A T U R E
R E V I E W
1. Understanding the Current State of Research

The initial stage of any scholarly endeavour is akin to

setting the stage for a grand performance; it
necessitates a comprehensive understanding of the
scene that has been set by previous acts. The
literature review, a foundation of academic research,
is the systematic approach employed to understand
the landscape of existing knowledge and to
appreciate the breadth of scholarly dialogue that has
shaped the topic at hand.

Reviewing the Scholarly Landscape

Embarking on a literature review is to embark on a

journey through the collective intellect of the field. It
requires an extensive and yet discerning exploration
of the vast array of studies, theories, and
perspectives that form the current state of research.
This review not only informs but also enlightens,
highlighting the achievements of past inquiry and the
terrains yet to be traversed.

Synthesizing Knowledge

The synthesis of information gathered from diverse

sources is an intricate process. It is like creating a
7
mosaic of knowledge, with each piece of research
acting as a tile in the bigger academic image. This
process requires careful analysis and a willingness to
consider different perspectives, to ensure that the
resulting overview is both precise and all-
encompassing.

Understanding Theories and Frameworks

A critical component of this phase is the

understanding of theories and frameworks that have
shaped the topic. It is about recognizing the
intellectual traditions and paradigms that have guided
past research, and identifying how they continue to
influence current and future studies.

Charting the Known and the Unknown

The objective is to chart a map of the known,

delineating the boundaries of current knowledge,
while simultaneously marking the uncharted
territories which is the questions that remain
unanswered, the problems that persist unresolved,
and the areas ripe for discovery.

8
The Role of Critical Analysis

In undertaking this pursuit, critical analysis is

indispensable. It is not merely a passive reception of
information but an active engagement with the
material. Researchers must question the
methodologies, dissect the arguments, and evaluate
the conclusions of existing research to fully
comprehend the state of the art in their field.

To understand the current state of research is to

know where one stands in the continuum of
knowledge. It is a foundational step that informs the
trajectory of the research journey, equipping scholars
with the insights necessary to contribute to the
ongoing conversation. It is this thorough grounding in
the literature that ensures the research is relevant,
significant, and poised to push the frontiers of
understanding.

2. Establishing Context and Relevance

In the grand tapestry of research, each thread of

inquiry is interwoven with those that have come
before it, creating a rich and textured narrative of
knowledge. Establishing context and relevance in a

9
literature review is the scholarly art of positioning
one’s work within this intricate fabric, highlighting its
connections to the broader academic discourse.

Situating the Study in the Scholarly Environment

The initial review into the existing literature serves not

only to inform but also to locate one's research within
the vast scholarly environment. This involves tracing
the intellectual lineage of concepts, situating
contemporary findings within historical trajectories,
and mapping the current research within the larger
constellation of scholarly works.

Illuminating the Pathways of Inquiry

A literature review transcends a mere summary of

previous works; it illuminates the pathways that have
led to the current state of understanding. It
demonstrates how each study contributes to the field,
it lays the groundwork for new research, establishing
its necessity and potential for impact.

10
Interrogating the Fabric of Knowledge

Contextualization: This process requires an

interrogation of the fabric of existing knowledge,
understanding not just the 'what' but the 'why' and
'how' of previous research.

Relevance: It also involves articulating the

relevance of new research questions in light of
what has been previously established, thus
asserting the significance of the study amidst the
sea of existing literature.

The Interplay of Continuity and Innovation

Continuity: The literature review establishes

continuity, demonstrating how the research is a
logical progression of/or a complementary
addition to existing scholarship.

Innovation: It also serves to highlight innovation,

showcasing how the study departs from/or
challenges established knowledge, thereby
adding a fresh perspective or addressing a critical
gap.

11
Creating a Narrative of Significance

The ultimate goal is to weave a narrative that

showcases the significance of the research. It is
about crafting a story that resonates with the
academic community, one that demonstrates the
researcher’s deep engagement and contribution to
their field.

When you carefully establish context and relevance,

it does not only justifies the importance of your study
but also ensures its integration into the ongoing
academic conversation. It is this careful positioning
that underscores the value of the research and its
potential to advance understanding within the
discipline.

3. Identifying Gaps and Questions

At the heart of academic progress lies the diligent

quest to uncover areas of ignorance which is the
voids in our collective understanding that invites
inquiry and exploration. The literature review stands
as the gateway to this quest, offering researchers a
lens through which the landscape of knowledge is
surveyed and scrutinized.

12
The Discovery of Gaps

A thorough literature review serves as a beacon,

illuminating the shadowy recesses where questions
remain unanswered. It is within these gaps that
opportunity resides as a chance to venture beyond
the known and contribute meaningfully to the tapestry
of scholarship.

Evaluating Existing Literature

The endeavour begins with a critical appraisal of

current literature. Each paper, each study, each
theory is dissected and examined however, not with
scepticism but with a constructive critical eye that
seeks to distinguish the known from the unknown.
Researchers must investigate the layers of
established knowledge, discerning not just the
content but the context, the methodology, the breadth
and depth of analysis. It is through this meticulous
examination that gaps are discerned. Sometimes
between the lines, at other times, in the grand
expanse of unaddressed questions.

13
Formulating Questions

From the fertile ground of these gaps sprout the

questions that will guide future research. Crafting
these queries is an art form therefore it demands
clarity, relevance, and foresight. The questions must
not only seek to fill the gaps but also challenge
existing paradigms, propose new angles of
examination, and open doors to further investigation.

Building a Structure for Inquiry

Upon these questions, we come up with our

hypotheses which is a tentative bridges between the
known and the unknown. A well-formed hypothesis is
both a prediction and a guide; it is a statement that
directs the research and predicates its outcomes. It
must be rooted in the literature, drawing from the rich
soil of past studies, yet it must reach out into the
unexplored, beckoning for verification or refutation.

Charting a Course

With the gaps identified and the questions poised, the

direction of the research becomes clear. This
newfound clarity charts a course that is both
ambitious and attainable. The identified gaps become
waypoints on the journey, guiding the research
through the uncharted waters of inquiry.
14
In identifying the gaps and questions, we find the
essence of research. It is a relentless pursuit of
knowledge that is forever pushing the boundaries of
what we know. A testament to the curious spirit of the
academic community and a commitment to the
unending quest for enlightenment.

4. Avoiding Duplication

In the scholarly pursuit of knowledge, the literature

review serves as an essential checkpoint for
originality. An exhaustive review of what has been
previously researched and published is fundamental,
allowing scholars to navigate the vast ocean of
academia without retracing the navigational paths of
their predecessors.

Conserving Academic Resources

The avoidance of duplication is not merely an

exercise in academic etiquette; it is an act of
conserving the most precious resources in research:
time and intellectual energy. Through systematically
identifying the explorations already undertaken,
researchers ensure that their efforts are channelled
toward unexplored territories, thereby enhancing the

15
collective body of knowledge with new insights and
discoveries.

Ensuring Incremental Advancements

Knowledge as a Continuum: Recognizing the

contributions of earlier studies is critical in
situating new research within the continuum of
knowledge. It fosters a scholarly dialogue that is
both progressive and cumulative.

Value Addition: Each academic endeavour should

aim to introduce a new dimension to what is
known, whether through the revelation of
overlooked details, the application of novel
methodologies, or the pursuit of unasked
questions.

5. Building a Theoretical Framework

The act of constructing a theoretical framework

through a literature review is akin to laying the keel of
a vessel that will navigate the academic seas. It is a
foundational endeavour that shapes the contours of
research, providing it with direction, structure, and a
lens through which to interpret findings.

16
Architecting the Research Scaffold

A literature review does more than acquaint one with

the prevailing winds of scholarship; it is the process
of erecting a scaffold upon which the entire research
project will rest. The framework serves as a
conceptual structure, outlining the key theories and
concepts that will underpin the research.

Synthesizing Foundational Theories

Integrating Core Concepts: This involves a

synthesis of core concepts and propositions that
have been established by previous scholarship,
creating a coherent structure that will support new
inquiry.

Theoretical Lineage: It is imperative to trace the

lineage of these theories, understanding their
genesis and the intellectual journey they have
traversed to reach their current form.

Formulating Theoretical Propositions

Constructing Propositions: From the synthesis of

literature emerges the construction of theoretical

17
 propositions. These are the building blocks of the
framework, each one a statement that posits a
relationship between concepts that the research
will explore.

Conceptual Relationships: The framework maps

out the conceptual relationships that the research
will test, challenging or reinforcing the
propositions with empirical evidence.

Critical Evaluation: The literature review provides

a critical evaluation of existing theories, assessing
their strengths, weaknesses, and applicability to
the new research context.

Identification of Theoretical Gaps: It also identifies

gaps within the theoretical landscape that the
research is poised to address, thereby
contributing to the evolution of theoretical
discourse.

The construction of a theoretical framework is an

exercise in intellectual craftsmanship. It requires the
researcher to engage deeply with the literature,
distilling it into a framework that not only supports but
also advances the study. It is through this

18
meticulously built framework that the research
acquires its academic rigour and theoretical
significance.

6. Methodological Insights

The exploration of methodology through a literature

review is a reflective journey into the heart of
research design. It is where scholars draw from the
collective wisdom of their field to forge the tools and
techniques that will shape their inquiry.

Gaining Methodological Clarity

A literature review is a scholarly expedition to gain

clarity on the various paths trodden by researchers
before. It is an examination of the choices made at
the crossroads of inquiry, offering insights into the
suitability, effectiveness, and potential limitations of
different research methods.

Surveying the Methodological Landscape

Assessment of Past Approaches: Scholars must

meticulously survey the methodological choices
of past research, discerning which approaches
yielded the most robust findings and which fell
short.
19
 Adaptation and Innovation: It involves not only an
adaptation of established methods but also an
innovation, as new questions may require new
methods or a novel combination of the old.

Informing Research Design

Informed Choices: The insights garnered from

this review inform the design of new research,
guiding choices about data collection, analysis,
and interpretation.

Best Practices: It also brings to light best

practices, ethical considerations, and potential
pitfalls, ensuring that the research design is both
sound and socially responsible.

The Interplay Between Theory and Method

Harmonizing Theory and Method: The review

ensures that the chosen methodology is in
harmony with the theoretical framework, providing
a consistent approach for answering the research
questions.

Refinement of Methods: Through critical

engagement with the literature, researchers refine

20
 their methods, shaping them to fit the specific
contours of their research questions and
objectives.

Methodological insights derived from a literature

review are the navigational instruments guiding the
research journey. They offer a lens through which the
research problem can be viewed, ensuring that the
methods chosen are best suited for exploring the
depths of the research questions posed. It is through
this meticulous preparation that the research is
positioned to yield findings of true academic worth.

7. Developing Critical Thinking

The literature review transcends the mere

accumulation of data; it is an exercise in intellectual
rigor, demanding the sharpening of critical faculties. It
is through this scholarly practice that one hones the
ability to not only absorb information but to dissect,
interrogate, and synthesize it, forging a mind that can
navigate the complexities of academic discourse with
acuity.

21
Cultivating Analytical Awareness

The process of engaging with the literature

necessitates an analytical awareness that can
penetrate beyond surface presentations to the
underlying arguments and assumptions. It requires
the discernment to differentiate between mere
opinion and evidence-backed assertion, between
robust methodology and methodological flaws.

Synthesis and Evaluation

Synthetic Vision: It is through the synthesis of

diverse perspectives that a researcher develops
the capacity to see the larger picture, to
understand how disparate pieces of knowledge
interlock within a broader intellectual puzzle.

Evaluative Judgement: The literature review

exercises the researcher’s evaluative judgement,
training them to assess the merit of scholarly
contributions critically and to weigh their
relevance to the research question at hand.

22
The Interrogative Process

Questioning as a Scholarly Imperative: Critical

thinking is rooted in the art of questioning. The
literature review propels scholars to ask incisive
questions about their field of study, prompting
them to challenge preconceptions and received
wisdom.

Developing the Researcher's Voice: Through this

interrogative process, researchers develop their
scholarly voice, learning to articulate their insights
with confidence and contributing original
perspectives to the academic dialogue.

Developing critical thinking through a literature review

is central to the maturation of a researcher. It is a vital
process that not only enhances individual scholarly
pursuits but also elevates the collective intellect of the
academic community. It ensures that the pursuit of
knowledge is not passive but dynamic, characterized
by a relentless quest for deeper understanding and
truth.

23
8. Establishing Academic Credibility

At the core of scholarly endeavours lies the pursuit of

academic credibility, a recognition of integrity and
rigor in research. The literature review serves as a
crucible where credibility is both tested and fortified. It
is through this meticulous process that researchers
demonstrate their commitment to the scholarly ethos,
grounding their work in the established canon of
academic knowledge.

Establishing Scholarly Legitimacy

The thorough examination of existing literature is

more than a procedural step; it is an affirmation of the
researcher’s dedication to the scholarly community. It
showcases a respect for the intellectual efforts of
predecessors and a recognition of the collective
building of knowledge over time.

Evidencing Depth of Knowledge

·Demonstration of Understanding: A comprehensive

literature review evidences a researcher’s deep
understanding of their field, showcasing an ability to
navigate the complexities of existing theories and
debates.
24
 Contextualization of Research: It situates the
research within a broader academic context,
illustrating how the study contributes to, extends,
or challenges the existing body of knowledge.

The Art of Scholarly Discourse

Engagement with the Literature: Effective

engagement with the literature is an art that
reflects the researcher’s ability to critically
analyse, synthesize, and draw insights from a
wide array of sources.

Communication of Expertise: The literature review

is also an opportunity for researchers to
communicate their expertise and understanding,
reinforcing their credibility within the academic
community.

Academic credibility, fostered through a well-

conducted literature review, is the foundation upon
which the integrity and value of research are built. It
is a vital aspect that underpins the researcher’s role
as a contributor to the collective endeavour of
knowledge advancement.

25
Now that we have revised the purpose and
importance of conducting a literature review, we will
explore the first process of conducting an effective,
credible and engaging literature review. As
mentioned earlier in this book, the first process of a
good literature review is an extensive literature
search which contributes to the rigorousness of the
literature review.

I have explained the purpose and importance of

conducting a literature review before presenting the
steps of conducting literature review in order to make
sure that we are focused on these while we are going
through the processes of conducting a literature
review.

26
 Part I
Literature
Search
F U N D A M E N T A L S
O F
L I T E R A T U R E
S E A R C H
The Conventional Method of Doing Literature
Search

The foundation of any academic research is the

literature search. It is a systematic and expansive
process that researchers undertake to identify and
collect all relevant information on a particular topic.
Actually, it seems quite strange to term this kind of
literature search as 'the conventional literature
search' as when I was actively doing literature search
for my PhD, this kind of search is deemed quite
revolutionary and cutting-edge.

However, with the advent of AI, this kind of search

seems way outdated as the AI assisted literature
search saves a lot of time. The conventional search is
a meticulous process that relies on the researcher's
understanding, patience, and analytical skills. It still
demands us to search for literatures from academic
journals and thesis repository. However, this can be
done online, rather than the traditional way of doing
literature search, that is doing physical searches in
physical libraries. The world is changing, even
libraries nowadays have stopped physical database
subscription and are investing on online databases.

29
These databases consume a large bulk of a library's
budget in order to provide students with a more
comprehensive and easy to access database. Now,
let me introduce you to some of these databases
starting from thesis or dissertation databases and
journal article databases.

Understanding Online Databases

An online database is a collection of related data

organized in a way that facilitates data searches and
information retrieval. For thesis and dissertation
writing, specific databases have been established to
store and share academic work. These databases
are crucial for conducting a comprehensive literature
search. Some of the most widely used databases for
theses and dissertation include:

Online Databases for Thesis and Dissertations

ProQuest Dissertations & Theses Global

This is an exhaustive collection of dissertations and

theses from around the world, spanning from 1743 to
the present day. It offers full text for graduate works
added since 1997, along with selected full text for
works written prior to 1997.

30
EThOS (Electronic Theses Online Service)

Run by the British Library, EThOS allows researchers

to access UK doctoral theses, providing a single point
of access to an extensive range of research.

DART-Europe E-theses Portal

This portal provides researchers with a single

European portal for the discovery of Electronic
Theses and Dissertations (ETDs).

Open Access Theses and Dissertations (OATD)

OATD.org aims to be the best possible resource for

finding open access graduate theses and
dissertations published around the world. I have a
fond memory of this database as my guiding thesis
was found using this database.

Networked Digital Library of Theses and

Dissertations (NDLTD)

This database was created by an international

organization dedicated to promote the adoption,
creation, use, dissemination, and preservation of
electronic theses and dissertations (ETDs).
31
Institutional Repositories

Many universities host their own institutional

repositories, which can be a rich source of
information. These can include theses, dissertations,
and other academic works by students and faculty.

Google Scholar

While not a specialized thesis and dissertation

database, Google Scholar provides a broad search of
scholarly literature across many disciplines and
sources. Please consider this as your database too
as sometimes what you are looking for might be here.

WorldCat

A global catalogue of library collections that allows

users to search for books, articles, dissertations, and
other media in libraries worldwide

Research Article Databases for Literature Search

Three of the most popular online database for

research articles are Scopus, Web of Science
(WOS), and Google Scholar. Each offers a unique set
of tools and resources for researchers.

32
Scopus

Scopus is akin to a gourmet food market, offering a

sophisticated and comprehensive collection of peer-
reviewed content. It is the largest abstract and
citation database of peer-reviewed literature,
delivering a broad overview of the world's research
output in the fields of science, technology, medicine,
social sciences, and arts and humanities. The
screenshot below illustrates what you would see
upon your first contact with the Scopus website.
Some of the features of the Scopus database are:

33
Broad Coverage

Scopus boasts extensive coverage, indexing a vast

array of journals and conference proceedings,
ensuring that researchers have access to a diverse
selection of sources.

Advanced Search Tools

With its advanced search capabilities, researchers

can create detailed and refined search strings that
yield highly targeted results.

Citation Tracking

Scopus allows for citation analysis, enabling scholars

to follow the trail of research through time and across
disciplines.

Analytic Tools

The database offers analytic features that can help

assess the impact of research, identify key opinion
leaders, and understand research trends.

34
Web of Science (WOS)

Web of Science can be compared to a specialty

culinary store that caters to the exquisite palate,
offering quality over quantity. WOS is a prestigious
citation indexing service that provides users with
powerful tools to access and analyse the impact of
research. Among the advantage of the WOS
database are:

Selective Indexing

It prides itself on a rigorous selection process,

ensuring that only high-impact and authoritative
journals are indexed.

Interdisciplinary Research

WOS facilitates cross-disciplinary research, making it

an ideal starting point for literature searches that
span multiple academic fields.

Citation Reports

It offers unique tools like citation reports and author

identifiers, which can help to map the influence of a
particular article or researcher.
35
Historical Reach

With coverage dating back to 1900, WOS is a rich

resource for historical literature searches.

Google Scholar

Google Scholar is the bustling city market of

academia where anyone can come to find a wide
array of research 'ingredients.' It is less selective than
Scopus and WOS, but its accessibility and simplicity
makes it a valuable tool for literature searches. The
screenshot below shows your first contact with the
Google Scholar site. Among the advantages of
Google Scholar are:

36
Wide Accessibility

Unlike its counterparts, Google Scholar is freely

accessible to anyone with internet access.

User-Friendly Interface

It offers a simple, user-friendly interface that is

familiar to anyone who has used Google's search
engine.

Broad Search Results

Google Scholar provides a broad search capability,

returning results from a wide range of disciplines and
formats, including theses, books, abstracts, and
articles from academic publishers, professional
societies, and universities.

Cited By Feature

Its 'Cited by' feature allows researchers to see how

many times and where a publication has been cited,
which is helpful for understanding its impact and for
snowballing.

37
Each of these databases has its own strengths and
specializations, much like how different shops cater
to various aspects of cooking. Scopus offers breadth
and depth with powerful analytics, WOS provides a
curated selection of impactful research, and Google
Scholar delivers wide accessibility and ease of use.
Researchers may visit one or all of these 'stores'
depending on their specific 'recipe' requirements for a
comprehensive literature search. I recommend using
all of them to ensure an extensive search of literature
has taken place.

The Importance of Keywords

The selection of keywords in conducting a literature

search is similar to choosing the right combination of
herbs and spices for a culinary masterpiece; they are
essential in defining the flavour and character of the
dish. In academic research, keywords are the
important terms that align with the core concepts of a
study, determining the relevance and precision of
search results. They that guide researchers through
the scholarly information, ensuring that the search is
focused, comprehensive, and fruitful. An example of
keywords can be seen in these screenshots of
articles below.

38
39
Just like a skilful chef expertly pairing flavours, a
researcher must carefully select and combine
keywords using Boolean operators and other search
techniques to extract the most pertinent studies from
vast databases. This meticulous process is the
foundation of a successful literature search, leading
to a well-informed and substantive literature review
that encapsulates the breadth and depth of existing
knowledge on a given topic.

The importance of keyword.

40
 D E V E L O P I N G
A N E F F E C T I V E
S E A R C H
S T R I N G
Developing an Effective Search String

A search string is a combination of keywords and

search operators (like AND, OR, NOT) that are used
to conduct a search in a database. Developing an
effective search string involves:

Keyword Identification

Begin by identifying the main concepts and variables

of your research question. Each concept may have
several terms associated with it. For example, the
concept 'adolescent' might also be found as
'teenager' or 'youth'.

Use of Boolean Operators

Combine your keywords using Boolean operators to

refine your search. For example, "adolescent OR
teenager OR youth" will retrieve records that contain
any of the synonyms.

Application of Search Limits

Most databases allow for setting limits to refine

search results further. These can include language,

42
date of publication, type of resource, and scholarly or
peer-reviewed status.

Phrase Searching and Truncation

Use quotation marks to search for exact phrases

("climate change") and truncation symbols to include
various word endings (e.g., child* to search for child,
children, childhood).

Examples of search strings

Scopus

This is a sample of an extensive search string for

literature search on the impact of social media on
mental health among adolescents:

TITLE-ABS-KEY ( ("social media" OR "online

networking" OR "digital communication") AND
("mental health" OR "psychological well-being" OR
"emotional distress") AND (adolescents OR
teenagers OR "young adults") ) AND ( LIMIT-TO (
DOCTYPE , "ar" ) OR LIMIT-TO ( DOCTYPE , "re" ) )
AND ( LIMIT-TO ( LANGUAGE , "English" ) ) AND (
EXCLUDE ( PUBYEAR , 2021 ) )

43
The screenshot below illustrate how this search string
can be used in Scopus database to search for
relevant literatures.

This search string uses TITLE-ABS-KEY to search for

keywords within the documents' title, abstract, and
keywords section. Boolean operators OR and AND
are used to expand or refine the search. The search
is limited to articles (ar) and reviews (re) published in
English, excluding the year 2021.

This particular search yielded 1567 related and

relevant documents for the search as it only excluded
the publications from the year 2021. For a more
relevant literature search, you can specify the
publication year to what is usually accepted within the
academic world, which is recent up to 5 years. An
example based on the time this book was written, a
literature search ranging from the year 2020 to 2024
would be highly appreciated.
44
However, extending the year of publication will yield
more result. This is very important too as reference
for seminal works (influential works), the guru's in the
field and the models and framework related to our
works.

There are many other search features in the Scopus

database and I have demonstrated them in a video
that can be accessed through the link below.

Scopus search features 1

Scopus search features 2

45
Web of Science (WOS)

This is a sample of an extensive search string for

investigating climate change effects on coral reefs:

TS=(("climate change" OR "global warming" OR

"temperature rise") AND ("coral reefs" OR "coral
bleaching" OR "marine ecosystems")) AND
Language: (English) AND Document Types: (Article
OR Review) NOT Year Published: (2021)

TS indicates that the search terms must be present in

the title, abstract, or keywords. The NOT operator is
used to exclude any documents published in the year
2021. Language and document type filters are
applied to refine the search further.

There are many other search features in the WOS

database and I have demonstrated them in a video
that can be accessed through the link below.

WOS search features

46
Google Scholar

This is a sample of an extensive search string for

exploring the use of artificial intelligence in
diagnosing cardiovascular diseases:

("artificial intelligence" OR "machine learning" OR

"deep learning") AND ("cardiovascular diseases" OR
"heart diseases" OR "cardiac diagnosis") -invasive -
surgery

In Google Scholar, parentheses are used to group

synonyms or related terms with OR to broaden the
search. The AND operator connects different
concepts relevant to the research. The minus sign (-)
is used before words like invasive and surgery to
exclude those terms from the search results.

The screenshot below illustrate the process of

literature search in Google Scholar using search
string.

47
48
 A I A I D E D
S E A R C H
S T R I N G
D E V E L O P M E N T
AI Aided Search String Development

The conventional way of constructing a search string

requires you to be skilful in using Boolean operators,
formulas and other elements needed for search string
construction. AI such as ChatGPT, Bard, Perplexity
and Claude can aid this process by providing you
suggestion of a search string for your intended
search.

An example of this can be seen in the following

screenshot which uses the topic 'mental health
awareness among children'. The prompt that I use is:

[please construct an extensive Scopus search string

for me on the topic 'mental health awareness among
children. limit the search to the year 2020-2023. I
want to use this to search for related articles in the
Scopus database]

50
Search String Development Using ChatGPT

This is me showing you that if you are not that skilful

in constructing a search string, you can ask ChatGPT
to suggest you a search string. You can copy the
search string as I have highlighted above. However, a
more reliable and effective result can be attained if
you can identify the keywords related to your topic
and search, and then provide them to ChatGPT.
Nonetheless, harnessing AI, particularly ChatGPT in
this process saves a lot of time and effort. You can

51
just edit and personalise the search string suggested
by ChatGPT. Learning on how to construct your own
search string can also be done using ChatGPT. All
you need to do is ask!

Even when you ask for a search string, ChatGPT still

take this opportunity to educate you on the
knowledge on search string construction as can be
seen in the following screenshot taken from ChatGPT
response for the same prompt given.

52
Other AIs such as Google Bard, Perplexity and
Claude can also be used for this purpose. Here is
how these AIs responded to the same prompt:

[please construct an extensive Scopus search string

for me on the topic 'mental health awareness among
children. limit the search to the year 2020-2023. I
want to use this to search for related articles in the
Scopus database]

Search string development using ChatGPT.

Search String Development Using Google Bard

53
Search String Development Using Perplexity

Search String Development Using Claude

54
The video tutorial for search string development using
AI can be accessed using the link below.

Search string development using Bard, Perplexity

and Claude.

You can even search for articles directly using these

AIs. Here is a sample search of literatures from
ChatGPT. Note that I am showing example from
ChatGPT Plus as the free version does not support
literature search.

This literature search was done using this prompt:

[find me relevant, current and real literatures on the

effectiveness of solar panel usage in Malaysia
setting]

55
Literature Search using ChatGPT

I used the LitMaps plugins for this purpose. You can

choose from a wide arrays of plugins available in
ChatGPT Plus for literature search such as PubMed,
Consensus Search, Scholar Ai, Science and many
more.

The following screenshots details how to activate

these plugins.

56
You need to choose plugins from the drop down
menu of your ChatGPT 4 as illustrated above. Next
you will need to choose Plugin Store as shown below.

Activating plugins.

57
After that, you need to search for the plugins that you
want to use in the plugin search box. If you are not
sure on what plugins to look for, you can search
using the keyword 'research' and you will have a lot
of plugins related to research and literature search to
choose from as can be seen below.

58
The video tutorial for literature search using ChatGPT
can be accessed using the link below.

Literature search using ChatGPT

You can also use other AIs for literature search such
as Bard, Perplexity and Claude. The examples are
given in the following screenshots.

Literature search using Bard

The same prompt for literature search was used and

Bard found 5 academic articles as the initial response
to the prompt.

59
I then pursue the search by asking Bard to arrange
the academic articles in a table format.

60
For easier reading, I prompted Bard to include
numbering for the items in the table.

61
I then prompted Bard to find more articles. Bard
found two more articles and the articles were
included in the table.

62
I then pursue this by asking Bard to provide the links
for these articles for easier retrieval.

63
The video tutorial for literature search using Bard can
be accessed using the link below.

Literature search using Bard

The same search can be done on Perplexity and

Claude.

64
Literature search on Perplexity

The same literature search using the same prompt

can also be done on Perplexity. The example can be
seen in the screenshots below.

65
The video tutorial for literature search using Bard can
be accessed using the link below.

Literature search using Perplexity

The same search done on Claude is portrayed below.

Literature search using Claude

The same prompt yielded this result on Claude.

The same request for links of the articles was made

in Claude. However, Claude cannot provide an
immediate access via links as can be seen in the
response.

66
I then prompted Claude to arrange the result in a
table format for better organization. It can do the
same thing as ChatGPT Plus, Bard and Perplexity.

67
The previous search result can be added simply by
requesting for addition of literatures as can be seen
below.

I then expanded the search towards South East Asia

region and Claude was able to locate and categorize
the academic articles found. This is very useful.
ChatGPT Plus, Bard and Perplexity can do this too.

68
The video tutorial for literature search using Claude
can be accessed using the link below.

Literature search using Claude

We have explored how AI tools like ChatGPT Plus,

Bard, Perplexity, and Claude can revolutionize the
way we conduct literature searches. Each tool offers
unique capabilities, and the results you get largely
depend on how creatively you use them. While
ChatGPT Plus requires a paid subscription to access
its advanced features, Bard, Perplexity, and Claude
offer similar services for free. This opens up a world
of possibilities for users, allowing them to access a
wide range of information and insights with ease. The

69
key takeaway is that the effectiveness of your
literature search using these AI tools is closely tied to
how you apply them, blending your own insights with
the information they provide.

However, it is crucial to remember that the

conventional method of literature search remains the
gold standard. Despite the efficiency and breadth of
information that AI tools offer, the conventional
approach to literature research is unmatched in terms
of reliability and accuracy. As we integrate AI into our
research processes, it is important to double-check
the information provided by these tools. AI can
significantly save time and enhance productivity, but
a balanced approach that combines AI assistance
with conventional research methods will yield the
most effective and time-efficient results. Remember,
AI is a powerful assistant, but the rigor and depth of
conventional literature search methods are
indispensable for thorough and accurate research.

70
As we integrate AI into our research processes, it is
important to double-check the information provided
by these tools. AI can significantly save time and
enhance productivity, but a balanced approach that
combines AI assistance with conventional research
methods will yield the most effective and time-
efficient results.

71
 T H E
C O N V E N T I O N A L
A N D A I
A S S I S T E D
S N O W B A L L I N G
S T R A T E G Y
The snowballing strategy in literature review is a
methodical approach that involves tracing the
research lineage of key publications by examining
their references and citations, thus gradually building
a comprehensive understanding of a particular field.
This technique can be undertaken manually, where
researchers meticulously sift through reference lists
to identify subsequent articles of interest, or it can be
significantly enhanced through AI assistance. Artificial
intelligence accelerates this process, employing
sophisticated algorithms to quickly parse through vast
amounts of bibliographic data, pinpointing interrelated
studies and extending the researcher’s reach within
the academic landscape with greater efficiency and
depth.

Manual Snowballing

Once relevant articles are identified, manual

snowballing can be employed. This involves
reviewing the reference lists of the articles to find
additional literatures that were not captured in the
initial search. It is a recursive process that can
uncover valuable resources which may have been
missed otherwise. An example of a reference section
in an academic article is shown below. You can look

73
for the cited papers in that particular research article
and search them to be used for your literature review.

The video tutorial for manual snowballing can be

accessed using the link below.

Manual snowballing

74
AI-Assisted Snowballing Strategy

Snowballing is a method used in literature review to

identify additional papers by examining the reference
lists of already known relevant papers. AI-assisted
snowballing leverages artificial intelligence to
streamline and enhance this process.

Advantages of AI-Assisted Snowballing

AI-assisted snowballing offers several advantages

over the manual process:

Speed

AI can process large volumes of data rapidly, going

through numerous documents in the time it would
take a researcher to review one.

Comprehensiveness

AI algorithms can identify patterns and connections

between papers that might not be immediately
evident to human researchers, leading to a more
comprehensive literature collection.

75
Unbiased Selection

While researchers may unconsciously be biased in

their selection of literature, AI can impartially evaluate
the relevance of references based on predefined
criteria, reducing selection bias.

How AI-Assisted Snowballing Works

Initial Collection

The process begins with the researcher identifying a

set of seed papers that are central to the topic. These
are fed into the AI system as the starting point for the
snowballing process.

AI Analysis

Using natural language processing (NLP) and

machine learning, the AI tool scans the references of
the seed papers. It evaluates the titles, abstracts, and
keywords of the cited works to determine their
relevance to the research topic.

76
Reference Extraction and Ranking

The AI tool extracts the bibliographic details of the

references and ranks them based on their relevance
to the research topic. It uses various metrics, such as
citation counts and the recency of publications.

Cross-Referencing and Expansion

The tool cross-references the extracted references

against multiple databases to find where they have
been cited elsewhere, expanding the snowballing
process beyond the initial seed papers.

Iterative Process

AI-assisted snowballing is iterative. With each round

of snowballing, the tool learns from the researcher’s
feedback on the relevance of identified papers,
refining its search and selection criteria for the next
iteration.

AI Tools for Snowballing

AI technology has revolutionized the process of

literature review, offering innovative tools that go
beyond traditional snowballing to assist researchers
in identifying relevant literature more effectively. Here
77
are some additional AI tools that serve the purpose of
snowballing in literature reviews.

Connected Papers

This is a unique tool that builds a graph of related

papers based on a single starting paper. It helps
researchers visualize a coherent narrative and
understand the academic genealogy of a paper by
showing prior work that led to it and subsequent work
that cites it. This visualization assists in snowballing
by clearly showing the lineage of research, making it
easier to discover relevant literature.

In order to use connected papers, you need to

register an account. The screenshot below details
this.

78
Connected papers offers free usage, however, they
limit the graph generation to only five graphs per
months. You can sign up for the plus version (paid
version) at as low as three (3) US dollars per month.

Next, you can either search for a relevant academic

paper related to your study, or provide connected
papers with the academic papers that you want to
snowball. You can do so in the 'search for a paper'
column. This is illustrated in the following screenshot.

You can now generate graphs from the papers. This

graph shows the papers connected to the papers that
you have searched. This is a special feature of
connected papers that can be very useful to you.

79
This is the automated snowballing being done for
you. You can now evaluate the papers shown in the
graphs and see how they are related to the paper that
you have determined as your 'primary academic
paper' in generating the graphs.

There are more features in the graphs output such as

the 'prior works' and 'derivative works'. The 'prior
works' feature is shown below.

80
When you click on the 'prior works' menu button,
connected papers will show you a list of papers that
were cited by your primary academic paper and the
connected papers within the same study scope and
discipline prior to the primary papers. These papers
can be downloaded as connected papers is also
integrated with semantic scholar.

The next function that is also important is the

'derivative works' feature. It allows you to snowball
papers that cited your primary papers and the recent
papers related to your study. This is illustrated in the
following screenshot.

81
When you integrate the use of connected papers in
your snowballing process, you save a lot of time and
can also yield effective results. The tutorial for
snowballing using connected papers is given below.
Click the link to access the video.

The video tutorial of snowballing for literatures using

connected papers can be accessed using the link
below.

Snowballing for literatures using connected papers

82
Litmaps

Litmaps takes a dynamic approach to literature

review, offering an evolving visual map that adjusts
as new papers are added. It is designed to keep
researchers updated on the latest developments in
their field. Litmaps goes beyond static lists, providing
an interactive experience that can make the
snowballing process more intuitive and less time-
consuming.

To start with Litmaps, you will need to first register an

account. This is free, however, if you wish to use the
advance features that Litmaps offers, you can
upgrade your account to a paid version. The interface
of Litmaps upon your first contact with it is shown in
the screenshot below.

83
To get started with Litmaps, you will need to first
discover the articles that you want to use as the
primary articles before you start the snowballing
process. The output for the snowballing process on
Litmaps is called the 'Seed Map'. In order to obtain an
effective seed map, Litmaps suggest users to input
five articles before generating a seed map. The
academic papers can also be uploaded using the
Import Articles button as can be seen in the following
screenshot.

84
The search result on Litmaps can be seen in the
screenshot below.

85
Once you have input articles (in this example five
articles), Litmaps then is ready to generate the seed
map for you in order to discover more related
academic papers connected to your primary articles.

You can then generate the seed map on relevant,

related and connected literatures in your field of
study. This is an automated snowballing done for
you. The search result includes search results from
the open based and subscribed database for
academic papers. This ensure a thorough literature
search utilising the snowball technique.

86
You can hover your cursor on top of the author's
name displayed on the seed map. Litmaps will then
show you the relation of this author with your primary
articles. It will also display you the details of that
academic article to ensure that you have a better
access in retrieving them.

87
Litmaps is now integrated with ChatGPT Plus as a
plugin feature. This transforms the literature review
process into a highly dynamic and interconnected
experience. This integration allows researchers to
visualize and navigate the complex landscape of
academic research, identifying key papers and the
connections between them, through an intuitive,
interactive map.

With ChatGPT Plus, users can leverage the

conversational AI's understanding of complex topics
to query the Litmaps database effectively. They can
receive tailored visual representations of how various
research articles are related. This can significantly

88
The video tutorial of snowballing for literatures using
Litmaps can be accessed using the link below.

Snowballing for literatures using Litmaps

Semantic Scholar

Semantic scholar is an AI-backed search engine that

provides a broad and depth exploration into academic
papers. It uses natural language processing and
machine learning to understand the context of papers
and their key contributions. One of its features
includes the ability to see which papers have cited a
particular article, thus enabling researchers to follow
the trail of scholarship from a single influential study
to newer research that has built upon it.

Research Rabbit

Research Rabbit is an innovative tool designed to

facilitate the snowballing method in academic
research. Snowballing is a technique where
researchers start with a core set of papers and then
expand their literature base by examining the
references within these papers, as well as the papers
that cite them. Research Rabbit excels in this process

89
by offering a visual and interactive platform where
researchers can easily track and visualize the
connections between academic papers.

When we input a few key articles, Research Rabbit

quickly maps out a network of related works,
highlighting how they are interconnected through
citations and references. This allows us to identify
influential papers, emerging trends, and gaps in the
literature more efficiently. The tool's intuitive interface
and advanced algorithms streamline the otherwise
time-consuming process of literature review, making
it easier for researchers to build a comprehensive
understanding of their field and ensuring that
important works are not overlooked in their research
journey.

The following screenshot shows your first contact

with Research Rabbit. You will need to register and
create an account with Research Rabbit in order to
use it. Research Rabbit is an open based AI which
means that you can use it for free.

90
Once you have registered to Research Rabbit, this is
the interface of your account. To start the process of
literature snowballing, you will need to create a
collection as illustrated in the following screenshot.

91
Next you will need to add academic articles to start
the process. This can be done by searching for
literatures or by uploading files as can be seen in the
following screenshot. For this demonstration purpose,
I have opted to search literatures on the topic of
'Academic writing using AI'.

Then you will need to choose a search provider as

can be seen in the next screenshot. You can choose
from either Biomedical & Life Sciences powered by
PubMed or All Subject Areas powered by Semantic
Scholar.

92
As can be seen in the following screenshot, Research
Rabbit have searched for the academic papers
related to my search. I then can add these papers
into my collection to be used as seed papers for the
snowballing process. You will get better result in
snowballing when you upload the papers that you
have read and determined are most relevant for your
research.

93
I have added 5 academic articles for the purpose of
this demonstration. Once these papers are already in
my collection, I can now generate multiple result from
these seed papers such as 'similar work', 'earlier
work', 'later work' and many others as can be seen in
the following screenshot.

94
I have chosen to show information on similar work.
The result displayed for similar work are 398
academic articles. Once I clicked on that, Research
Rabbit generates a map showing similar works as
compared to my seed papers and illustrates how they
are connected. This can be seen in the following
screenshot.

95
There are many options given on how you want these
information to be displayed. The previous screenshot
illustrated the default setting showing a network type
graph. In the following screenshot, I demonstrated
what a timeline graph type would look like.

96
The academic papers displayed on the graph can be
selected and information of the academic papers will
be displayed. You can even download the academic
paper.

97
These are some of the features of Research Rabbit.
There are many more features within Research
Rabbit and you can explore them on your own. Next,
I would like to introduce another AI that can be useful
for the snowballing of literatures, Elicit.

The video tutorial of snowballing for literatures using

Research Rabbit can be accessed using the link
below.

Snowballing for literatures using Research Rabbit

Next, I would like to introduce another AI that can be

useful for the snowballing of literatures, Elicit.

98
Elicit

Elicit, another cutting-edge research tool, offers yet

another unique approach to the snowballing method
in literature review. The following screenshot shows
what to expect on your first contact with Elicit. As
other AIs, you also need to register for an account
before you can start using Elicit.

Elicit can be used free with limited features or you

can also subscribe to its Plus account at 10USD per
month.

99
With Elicit, you will need to ask a research question
to Elicit so that it can start finding papers for you.

100
For demonstration purpose, I asked the question:
'How are AIs being used for research writing?'

Elicit found 4 papers and constructed a summary for

the search. It summarises the abstract of these
papers.

101
You can click on the citations in the summary and get
access to the papers that you want.

The video tutorial of snowballing for literatures using

Elicit can be accessed using the link below.

Snowballing for literatures using Elicit

Next, I will demonstrate how Scite can be used for

literature snowballing.

102
Scite

Scite is an advanced citation analysis tool that not

only allows researchers to see how a scientific paper
has been cited but also provides the context of each
citation. It differentiates whether the citing articles
provide supporting or contrasting evidence for the
cited claims. This detailed citation context aids in the
snowballing process by helping researchers to
quickly assess the impact and relevance of the
literature.

Up until the time this book was written, Scite only

offers a 7 day free trial. After you have exhausted this
trial, you will need to subscribe to Scite in order to
use it. The subscription is 7 USD per month. The
screenshot tutorials on conducting literature search
with Scite can be seen in the following screenshots.

To use Scite, just like other AI tools, you will need to

register an account for it. However, unlike other tools,
Scite does not offer a continuous free usage at a
limited feature accessibility. You will need to
subscribe in order to continue to use Scite. The
screenshot below shows what to expect upon your
first contact with Scite.

103
Once you search for a particular topic or academic
papers, the search result will display the related
academic articles based on your search. However,
Scite offers a more detailed information compared
with other AI tools as it also provides the exact
quotation related to your search. This can be seen in
the following screenshots.

104
105
Another advance feature of Scite is that on top of
abundant drop down menus regarding the information
on your literature search, it also provide the types of
citation and the degree of agreement of other
academic research article with your primary article or
search topic. An example of this can be seen in the
following screenshot.

You can choose from three options. You can either

display literatures that either support, contrast or only
mention your primary articles. This is a very useful
feature for students and researchers as it will speed
up the process of literature search. On top of that, by
giving the exact quotation for each related searches,
we can directly be guided to the important details that
once will need hours, if not days, of careful reading.

106
This does not mean that you do not have to read the
academic papers related to your primary papers. I
think I do not have to mention that you need to do
due diligence to the research process as to maintain
integrity and quality of your output. However, using
Scite definitely will save you a lot of time and energy.

The video tutorial of snowballing for literatures using

Scite can be accessed using the link below.

Snowballing for literatures using Scite

Mendeley

Mendeley can also be used to search for literatures.

While primarily used for reference management and
PDF organization, its recommendation system can be
seen as a snowballing aid. Mendeley analyses the
user's library and reading habits to suggest new and
relevant papers, essentially automating part of the
snowballing process by bringing forward literature
that has thematic relevance to the user's existing
collection.

107
Iris.ai

This is an AI-powered science assistant that helps

researchers explore interconnected research ideas
and find relevant studies. It uses a combination of
machine learning algorithms and semantic
technology to map out the scientific landscape related
to a particular topic or paper. Iris.ai analyses the
content of a seed paper and provides a visual
exploration map that links to other research papers,
grouped by their relevance and interconnections,
facilitating a form of AI-driven snowballing.

There are a lot of other AIs for literature search which

are not listed here. It is up to you to decide which one
would best suit you. The AIs that I have shared is the
ones that I feel is most helpful for me based on my
experience. I recommend you to explore other AIs
and choose which one that caters to your need the
most. However, please do not be to focused on trying
all of the option as the main focus should be to
choose one that you are most comfortable with and
gets the job done.

Integrating these AI tools into the literature review

process can significantly enhance the ability to

108
identify relevant research and ensure that a literature
review is as comprehensive and up-to-date as
possible. Each tool brings a different strength to the
table, whether it is visual mapping, contextual citation
analysis, or personalized recommendations, thereby
contributing to a more robust and interconnected
literature review.

AI-assisted snowballing is a powerful strategy for

conducting thorough literature reviews. It enhances
the efficiency and effectiveness of the traditional
snowballing method, allowing researchers to uncover
a wider array of relevant literature with reduced effort
and increased accuracy. When integrated properly,
AI-assisted snowballing can significantly advance the
depth and quality of a literature review.

Conclusion

The journey through the fundamentals of literature

search has laid a solid foundation for understanding
the comprehensive process of identifying and
collecting academic resources. The conventional
methods, which is also technologically advanced,
have stood the test of time, relying on the diligent

109
efforts of researchers to navigate through vast online
databases and meticulously develop search strings
that unearth valuable scholarly works.

Databases like ProQuest, EThOS, and the

Networked Digital Library of Theses and
Dissertations serve as crucial gateways to extensive
research documents, while Google Scholar and
WorldCat offer broader platforms for academic
exploration. The art of crafting an effective search
string and combining keyword identification with
Boolean operator remains a critical skill for any
academic endeavour.

We have also ventured into the vast capabilities of AI-

assisted literature search, which promises to
revolutionize the traditional methodologies. AI tools
like Scite, Connected papers, and Litmaps, bring
speed, precision, and an unbiased approach to the
literature search process. They simplify the once
labour-intensive task of snowballing through iterative
processes and visual mapping, ensuring that
researchers can build upon a solid base of existing
knowledge with ease and efficiency.

110
In this chapter, we have seen how AI is not replacing
the researcher but rather augmenting their
capabilities, allowing for more time to be spent on
analysis and interpretation rather than on the initial
search. As we concluded our exploration of AI-
assisted literature search, it became clear that the
future of literature search is one where human insight
and artificial intelligence work in tandem to push the
boundaries of knowledge and discovery.

We must understand the importance of a well-

conducted literature search. It is the foundation which
all subsequent research efforts are built, and with the
integration of AI tools, the potential to expand our
understanding and contribute meaningful insights to
our fields of study has never been greater.

In conclusion, the literature search is the beginning of

a process where one gathers the best 'ingredients' to
create a 'dish' that is thoughtful, well-prepared, and
ultimately contributes to a greater understanding of
the research 'cuisine.' The more relevant and high-
quality the 'ingredients,' the richer and more
'delicious' the final 'dish', that is the literature review,
will be.

111
R E T R I E V I N G
L I T E R A T U R E
Accessing Academic Literature

In the vast culinary world of academic research,

retrieving literature is much like sourcing the right
ingredients for your kitchen. There are two primary
markets where these ingredients can be found: open-
access and subscription-based databases. Open-
access databases are akin to a farmer's market
where produce is freely available to anyone who
visits, representing the democratization of knowledge.
In contrast, subscription-based databases are more
like exclusive membership clubs where only
registered members can access their premium stock.

When retrieving literature, researchers must navigate

through these diverse markets, employing various
methods to obtain the necessary resources. They
may use direct search techniques on specific
databases, employ reference management tools to
organize their findings, or reach out to interlibrary
loan services to access documents not readily
available to them.

Yet, just as ethical sourcing is essential in the

culinary world, so too is it in academic research.
Researchers must respect copyright laws and

113
licensing agreements when accessing literature. The
use of certain resources like Scihub, Zlibrary, and
Libgen, which operate in legally grey areas, is a
contentious issue that requires careful consideration.
Scholars must balance their need for information with
the legal and ethical implications of how they obtain
it.

AI in Retrieving Literature

The emergence of AI in the kitchen of academic

research has introduced innovative gadgets and tools
that make the retrieval process more efficient. Tools
like ChatGPT Plus and SCite are the high-tech
appliances that help researchers sift through the vast
amount of information with ease, identifying and
retrieving the most relevant papers, articles, and
publications.

ChatGPT Plus acts as a personal sous-chef,

equipped with plugins such as Litmaps and
Connected Papers, which assist in not only finding
the right ingredients but also understanding how they
fit into the larger recipe of existing research. SCite, on
the other hand, is like a sophisticated food processor
that can take a paper and break it down into its

114
citations, analysing how each reference is used and
providing insights into its reliability and impact.

In addition to these tools, databases like Semantic

Scholar and Dimensions.AI are the specialty stores
providing niche ingredients that may not be found in
more general markets. Semantic Scholar utilizes AI to
cut through the noise, offering a more focused way to
discover relevant literature. Meanwhile,
Dimensions.AI offers a broad scope, similar to a
wholesaler, providing vast quantities of research
papers from various disciplines.

These AI tools and databases are transforming the

landscape of literature retrieval, making it possible to
gather a more relevant and comprehensive set of
'ingredients' for the 'dish' that is the literature review.
With these resources, researchers can ensure that
their scholarly 'meals' are prepared with the finest
'produce' available, satisfying the 'appetites' of their
academic 'diners'.

115
 Part II
Literature
Management
O R G A N I Z I N G
L I T E R A T U R E
W I T H
M E N D E L E Y
The process of organizing literature is a critical step
in academic research, similar to a meticulous filing
system that must be both functional and accessible.
With the advent of AI tools, the traditional landscape
of literature organization has evolved significantly. AI
provides a powerful suite of capabilities that can
automate, enhance, and streamline the organization
of research materials.

Organizing Literature Using Mendeley

Mendeley stands out as a pioneering reference

management tool that assists researchers in
organizing their literature collections. It acts as a
personal library where scholars can import
references, documents, and notes, organizing them
into folders and sub-folders for easy retrieval.
Mendeley’s AI features offer smart recommendations
and can automatically generate bibliographies and
citations in various styles, simplifying the meticulous
task of reference management.

The process of managing literature with Mendeley

involves several key steps. Below is a detailed guide
on how to use Mendeley for literature management:

118
Step-by-Step Process of Literature Management
with Mendeley

Step 1: Creating a Mendeley Account

Begin by signing up for a Mendeley account on

their website.
Once registered, download the Mendeley
Desktop application and sign in.

Step 2: Importing Documents

Click on the “Add” button on the Mendeley

Desktop to import documents.
You can import documents by selecting files from
your computer, dragging and dropping files into
Mendeley, or importing documents directly from
other databases.

119
You can also create a watch folder for easier
importing of documents into you Mendeley. A watch
folder means that any paper that were included in
your literature folder (on your computer) will be
automatically added and updated into you Mendeley
account. You no longer need to manually add the
documents into your Mendeley.

The video tutorial of importing documents to

Mendeley can be accessed using the link below.

Importing documents to Mendeley

120
Once a document was added into Mendeley,
Mendeley will automatically extract metadata (like
titles, authors, and publication details) to create your
library entries as can be seen in the following
screenshot.

The metadata are displayed on the right section of

your Mendeley. These metadata need to be correct
as these information will be used to create citation
and bibliographies by Mendeley. The metadata
includes author(s) name, journal, year, issue, pages
and others.

The video tutorial of identifying metadata information

in Mendeley can be accessed using the link below.

Identifying metadata information in Mendeley

121
Step 3: Organizing Your Library

Create folders within Mendeley to categorize your

documents according to your research themes or
chapters.
Simply drag and drop documents into these
folders to keep your library organized.
Use the “Tags” feature to tag articles with
keywords for easier retrieval.

The video tutorial of creating folders in Mendeley can

be accessed using the link below.

Creating folders in Mendeley

122
Step 4: Reading and Annotating

Open documents within Mendeley to read and

annotate them.
Highlight text, add notes, and create bookmarks
that will be saved directly within your library.
These annotations can be accessed later for
review and are searchable within the application.

The video tutorial of reading, highlighting and

inserting notes in Mendeley can be accessed using
the link below.

Reading, highlighting and inserting notes in

Mendeley.

123
Step 5: Citing While You Write

Install the Mendeley Citation Plugin for your word

processor.
As you write, use the plugin to search for
references in your Mendeley library and insert
citations into your document.
The plugin will automatically create a bibliography
from your cited references.

Citing while you write.

124
Step 6: Collaborating with Others

Share folders with colleagues or your supervisor

to collaborate on literature.
You can set up private groups where members
can share annotations and notes on shared
documents.

125
Step 7: Syncing Across Devices

Ensure that your work is backed up and

accessible across devices by syncing your library.
Click the “Sync” button to update your Mendeley
Web account with any changes made on
Mendeley Desktop.

Step 8: Staying Up to Date

Set up alerts and use the Mendeley Web platform

to discover new research articles and data.
Mendeley will recommend articles based on your
library content and your reading history.

126
Step 9: Generating Bibliographies

Select references in your library and use

Mendeley to generate a bibliography.
You can customize the format and citation style
according to your requirements.

The video tutorial of generating bibliographies using

Mendeley can be accessed using the link below.

Generating bibliographies using Mendeley

127
Step 10: Maintaining the Library

Regularly update your Mendeley library by adding

new research and removing outdated references.
Check for duplicate entries and update metadata
when necessary to keep your library accurate and
current.

128
D E V E L O P I N G
T H E
L I T E R A T U R E
M A T R I X
Developing the Literature Matrix

Creating a literature matrix is similar to preparing a

complex and hearty meal, where each component
must be carefully prepared and assembled. Just as a
chef curates ingredients, categorizing them by type,
flavour profiles, and cooking times, a researcher
develops a literature matrix to organize their findings
into a digestible format. The matrix acts as the
kitchen countertop where ingredients, which is the
themes, methodologies, findings, and key debates of
the literature, are chopped, sifted, and laid out.

The process begins with selecting the freshest

'ingredients', that is the most relevant and current
research. These are then 'marinated' with AI tools
that extract the 'juices' of the content, such as the
study design, population, outcomes, and conclusions.
This automation is like a food processor that does the
preliminary prep work, allowing the researcher-chef to
focus on the 'cooking' that is the analysis and
synthesis of the information.

As the 'dish' comes together, the researcher

compares and contrasts 'flavours' (information),
identifying which 'ingredients' complement each other
and which provide a necessary contrast, highlighting

130
patterns or trends much like a taste test reveals what
a dish might be missing. The literature matrix
becomes a recipe card, with clear instructions on how
each 'ingredient' contributes to the overall 'meal'
which is the literature review.

Following this structured 'culinary' process, the

researcher-chef ensures that their final 'dish', the
written literature review, is both comprehensive and
savoury, offering a balanced 'taste' of the available
research and leaving readers with a clear
understanding of the academic 'meal' that has been
presented to them.

Steps of Developing A Literature Matrix

The steps of developing a literature matrix are

detailed below.

Step 1: Define Your Objectives

Before you begin filling in a literature matrix, clearly

define what you aim to achieve with it. Determine the
key themes, variables, or research questions that you
are exploring. This will guide what information you
extract from the articles.

131
Step 2: Choose a Format for Your Matrix

Decide on a format for your literature matrix. It can be

a spreadsheet, table, or database, depending on your
preference and the volume of literature. Spreadsheet
applications like Microsoft Excel or Google Sheets
are commonly used because they are flexible and
allow for easy sorting and editing.

Step 3: Identify Categories and Headings

Based on your objectives, create categories and

headings for your matrix. Common headings include:

Author(s) and Year

Title
Research Questions/Hypotheses
Methodology (including research design,
participants, tools, and techniques)
Key Findings/Results
Theoretical Framework
Strengths and Limitations
Contributions to the Field
Notes (for additional comments or thoughts)

132
Sample of a literature review matrix can be seen in
the following table.

Literature matrix table template.

Step 4: Populate the Matrix

Begin to fill in the matrix with information from your

sources. Read each article, book, or study and
extract the relevant information for each category.
This step can be time-consuming but is critical for
building a comprehensive matrix.

133
Step 5: Analyse and Synthesize Information

Once your matrix is populated, start analysing the

collected data. Look for patterns, differences, and
common findings across the different studies.
Synthesize the information to draw broader
conclusions or to identify where there is a lack of
evidence or conflicting results.

Step 6: Update Continuously

As your literature review progresses, continuously

update the matrix with new sources or findings. A
literature matrix is a living document that evolves as
you go deeper into your research.

Step 7: Use the Matrix to Write Your Review

Use the synthesized information from your matrix as

a foundation for writing your literature review. The
matrix will help you structure your review, ensuring
that you cover all relevant aspects and provide a
critical analysis of the existing literature.

134
Step 8: Refine and Reorganize As Necessary

As your understanding of the topic deepens, you may

need to refine or reorganize your matrix. This could
include adding new headings, merging categories, or
re-evaluating the relevance of certain studies.

Step 9: Draw Conclusions

Finally, use your literature matrix to draw conclusions

about the state of the research on your topic. Identify
where consensus exists, where debates or
discrepancies arise, and where further research is
needed.

A well-developed literature matrix not only aids in

writing a literature review but also serves as a
valuable tool for discussions with supervisors or
colleagues, as well as for preparing for
comprehensive exams or research proposals. It
provides a clear, visual representation of the research
landscape and ensures that you have a firm grasp of
the literature in your field.

Conventional way of literature matrix.

135
A I A S S I S T E D
L I T E R A T U R E
M A T R I X
This chapter explores the innovative use of AI, with a
specific focus on how it can be leveraged to construct
a literature matrix which is a powerful tool for
organizing, analysing, and synthesizing research
findings. The creation of a literature matrix,
traditionally a manual and time-consuming task, can
be significantly enhanced with the capabilities of AI,
resulting in a more efficient and insightful process.

As we continue our journey in literature management,

I will provide practical examples and screenshots as
an idea on how ChatGPT can be used to generate
literature matrix. These examples demonstrate the
seamless integration of AI in enhancing the quality
and depth of literature reviews.

AI Aided Literature Matrix Using ChatGPT

I begin my request to ChatGPT with a simple prompt:

[build me a template of literature matrix for a PhD

student. Build it in a table format in Ms Word]

137
138
Notice the spelling error in the prompt from the
screenshot? This prompts was written to emulate real
case use of ChatGPT. Therefore, I did not edit my
prompt (instruction to ChatGPT) to showcase that
you do not need to edit prompts when instructing
ChatGPT to do a task for you. If your prompt contains
simple punctuation, spelling and grammatical
mistakes, ChatGPT can still detect what you actually
mean and act upon the prompt.

In this particular prompt, I have asked ChatGPT to

come up with a literature matrix template in Ms Word.
ChatGPT responded by suggesting the structure for
my literature matrix template that is suitable for a PhD
student (since I requested that). I have also creatively
asked ChatGPT to produce the template in Ms Word
format. You can also request it in an Ms Excel format,
or just a table in the usual ChatGPT response.

I did not edit the structure of the template as my

purpose is to show you how this can be done using
ChatGPT. However, you can edit and personalise
your literature matrix template according to your
needs. However, I edited the format of the template in
Ms Word such as the layout (changed to landscape)
and margin (changed to narrow) of the document for

139
a more effective and aesthetic literature review
template. I then uploaded the personalised literature
matrix template and instruct ChatGPT to use the
template for my upcoming request as can be seen in
the subsequent prompt.

[use this template from now on for literature matrix

template]

Next, I uploaded a literature (an academic article) and

prompted ChatGPT to analyse it and update the
template using information from the academic article.
The prompt is:

[analyse this article and fill in the previous literature

template with relevant information from this article].

140
After that, I then uploaded more papers and gave a
more simplified prompt that is:

[analyse these and fill in the template].

141
ChatGPT then analysed the article one by one and
updated the literature review matrix using information
from the analysed article, before doing the same task
for the other articles. This can be seen in the
following screenshot.

Within seconds I now have three academic papers

analysed and arranged into my literature review
matrix. As good as it sounds, the information that
ChatGPT extracted and included in the literature
matrix needs to be verified and personalised based
on your reading and understanding of the academic
article. You cannot just simply trust and rely on
ChatGPT to do your literature matrix for you.

142
The video tutorial of developing a literature matrix
using ChatGPT
can be accessed using the link below.

Literature matrix template using ChatGPT

Literature matrix development using ChatGPT

I risk sounding like a broken record, however, the

manual and conventional way of doing things (in this
case the literature matrix) is the most accurate way of
dealing with this process. The conventional way
ensures the reliability of the information and promotes
knowledge advancement on your part. This is
especially more important if you are a student. Both
researchers and students need to understand their
literatures before extracting them into literature matrix
to ensure that the information are useful for them.

However, as I have already illustrated, the use of

ChatGPT can aid this process to make it faster.
Combining both the conventional method and the AI
assisted method would be the most effective way of
managing your literature through a comprehensive
and useful literature matrix.

143
I risk sounding like a broken record, however, the
manual and conventional way of doing things (in this
case the literature matrix) is the most accurate way of
dealing with this process. The conventional way
ensures the reliability of the information and promotes
knowledge advancement on your part. This is
especially more important if you are a student. Both
researchers and students need to understand their
literatures before extracting them into literature matrix
to ensure that the information are useful for them.

144
The Role of AI in Managing and Organizing
Literature

AI's role in managing and organizing literature is

transformative. AI-powered tools can classify
literature based on content, suggest tags for themes,
and even highlight trends and commonalities across
different studies. They can learn from the user’s
interactions and progressively provide more
personalized and relevant content, optimizing the
organization process over time.

Detailed Exploration of AI-based Literature

Management Tools

Beyond Mendeley, there are other AI-based tools that

specialize in literature management. Tools like Zotero
use AI to provide insights into the metadata of articles
for better sorting and retrieval. Others, like
ReadCube, employ AI to enhance PDF reading
experiences, allowing for annotations and references
to be automatically sorted and analysed.

145
Strategies for Efficient Literature Categorization
Using AI

Efficient literature categorization involves not just

sorting by topic, but also understanding the context
and subtle information within the research. AI assists
in this by employing natural language processing to
interpret the content and suggest categorizations that
might not be immediately obvious to the human eye.
These strategies can include clustering similar works,
tagging literature with customized labels, and even
predicting future trends by analysing current
categorization patterns.

In conclusion, AI tools have become indispensable

allies in the quest to tame the ever-growing jungle of
academic literature. They provide a level of
organization that is meticulous, customizable, and
continuously evolving, allowing scholars to focus
more on analysis and less on the mechanics of
organization. As we progress through this chapter,
we will experiment into each of these aspects in
greater detail, exploring the practical applications and
benefits of AI in organizing literature.

146
 Part III
Literature
Comprehension
 R E A D I N G
S K I L L S F O R
L I T E R A T U R E
R E V I E W
When you start a literature review, you are setting off
on a journey through a vast landscape of information.
This chapter is your guide, teaching you how to read
effectively to make sure your literature review is
successful and not too overwhelming.

Literature Review Reading Techniques

Pre-reading Strategies

It is important to set your goals before you begin.

Think about what you want to find out from your
reading. Scan through the articles first, looking at
titles and subtitles to help you focus on the important
parts.

Active Reading Techniques

Active reading means really engaging with the text.

Ask questions, make notes, and highlight key points.
This will help you to remember and understand the
material better.

149
Critical Reading

Read with a careful eye, questioning the information

and the evidence given. Look out for any bias the
author might have. This will help you decide whether
you can trust the information.

Speed Reading vs. Deep Reading

There will be times when you read quickly to get the

gist and times when you need to slow down to
understand complex ideas. Knowing when to skim
and when to examine into the reading material is key.

Analytical Reading

Focus on the details of how the studies were done

and what the findings were. Think of yourself as a
detective trying to understand how all the parts fit
together.

Reading for Synthesis

Putting together information from different sources is

like completing a puzzle. You are combining many
small pieces to see the complete picture.

150
Managing Overwhelming Volume of Literature

With so much to read, you need to choose what is

most important. Use summaries and abstracts to
decide whether you should read the full article. Refer
to citations to find other significant works.

Utilizing Technology and Tools

There are many digital tools available to help you

read and organize your findings. They can save you
time and help keep your work in order.

Enhancing Comprehension and Retention

To make sure you remember the material, review

your notes later and discuss what you have learned
with others. This is like rehearsing a story so it stays
with you.

Reading for a literature review is a skill that improves

with practice. Use these tips as you work, and you
will find your comprehension and writing improve.

151
 A I A S S I S T E D
L I T E R A T U R E
C O M P R E H E N S I O N
This chapter moves on from the reading skills we
talked about before. Now, we are looking at how AI,
which is short for artificial intelligence, helps us
understand long and difficult texts. AI works like a
helpful guide that can quickly read through lots of
complicated writing. It makes everything clearer and
easier to understand. In this part of the book, we will
explain how AI is really useful for making sense of
complex ideas in what you read.

We will also see how AI can help summarize big

pieces of writing. It can find the most important ideas
fast. This is a big help for people doing research
because it means they can learn a lot without reading
every single word. AI can spot patterns and main
themes, and it can show where more information
might be needed. Using AI can help researchers
have a good understanding of their area of study, and
it can make the process of reviewing literature much
quicker.

We will explore how specific tools like ChatGPT and

ChatPDF can be used to improve our understanding
of scholarly texts. ChatGPT can act as a
conversational partner, helping you to ask and
answer questions, and get explanations in real-time,

153
which makes tough concepts easier to grasp.
ChatPDF, on the other hand, allows you to dive into
documents, using AI to pull out key information,
summarize texts, and even identify where the
discussion in a field is lacking. Using these tools, you
can get a stronger grip on your literature review,
identifying the main ideas and themes with the help of
technology. I will guide you through practical ways to
use ChatGPT and ChatPDF to enhance your
research process, making it faster and more
thorough.

Using ChatGPT for Literature Reading and

Comprehension

There are four methods in using ChatGPT for

literature reading and comprehension such as using
ChatGPT 3.5 (free version) and using ChatGPT Plus
(paid version). In using ChatGPT 4 (paid version),
there are three options of conducting literature
reading and comprehension such as using Plugins,
Advanced Data Analysis and extensions of ChatGPT
4 in Google Chrome.

154
The first method involves using the freely available
ChatGPT 3.5. In this approach, you must manually
select portions of an article (from the entire text) and
paste it into the ChatGPT interface. Following this,
you can engage with ChatGPT by posing questions
related to the pasted content. This interactive process
allows for a deeper understanding of the material, as
ChatGPT helps clarify complex concepts, explain
methodologies, and provide definitions or summaries.

Moving to a more integrated experience, the second

method makes use of PDF Reader plugins within
ChatGPT Plus, also known as ChatGPT 4. This
version offers a variety of PDF reader plugins that
empower users to 'chat' directly with the PDF
document. With queries and questions, you can
navigate through the document and receive targeted
responses, essentially holding a conversation with
the text. This not only streamlines the comprehension
process but also makes it more engaging and
efficient.

The third method leverages Advanced Data Analysis

features in ChatGPT 4. With this powerful tool, users
simply upload the article into the ChatGPT platform,
where the advanced data analysis capabilities take

155
over. The system thoroughly analyses the content of
the article, after which users can ask questions. The
AI provides insights based on its analysis, offering
users a comprehensive understanding of the text
without the need for manual intervention. This
method is particularly useful for quickly grasping the
key points and subtle details of complex literature.

The fourth method is the use of extensions of

ChatGPT 4 in Google Chrome for literature reading
and comprehension. This method represents a
seamless integration of AI into your everyday web
browsing experience. Utilising Google Chrome
extension, ChatGPT 4 becomes an on-hand assistant
as you navigate through various online literary texts.
This approach is incredibly user-friendly and efficient.

You can simply highlight text on any webpage, and

the extension will provide you with an instant
analysis, interpretation, or summary directly within
your browser. This method is particularly useful for
those who frequently read online articles, e-books, or
digital libraries. It allows for real-time comprehension
assistance, making the process of understanding and
analysing literature more accessible and immediate.
This integration of ChatGPT 4 into your browser

156
transforms it into a powerful tool that enhances your
reading experience, providing insights and
clarifications at the click of a button.

Literature Comprehension Using ChatGPT 3.5

Using ChatGPT 3.5 for literature comprehension

offers a valuable advantage: it is free. This
accessibility means that anyone with an internet
connection can start enhancing their understanding of
scholarly texts without worrying about subscription
costs. However, there are trade-offs to consider.
Since ChatGPT 3.5 does not directly integrate with
PDFs or other document formats, users must put in
extra effort to copy and paste sections of the text into
the chat.

This can be time-consuming and may slow down the

overall reading process. Additionally, it requires more
active work from the user to ensure that the pasted
content is in a manageable format for the AI to
interpret accurately. Despite these drawbacks,
ChatGPT 3.5 remains a useful tool for those looking
to engage with complex literature in a budget-friendly
way.

157
The screenshots below will detail how this can be
done. I start with prompting ChatGPT 3.5 with the
following prompt:

[I will provide you with an article. Please read and

understand the article. After that, I will ask you
questions regarding the articles. I will paste the
articles in batches due to word limit restrictions.
Please only respond with a YES after I have pasted
the articles, and then I will proceed with the
processes until it is done. The article :(paste parts of
your article here)]

I then pasted the copied article's section as pasting

the whole document would exceed the limit of words
that ChatGPT can process per single prompt.
Usually, for an academic article, you need to segment
it into three sections for this manual delivery to
ChatGPT. The screenshot below illustrates this:

158
ChatGPT then responds with only a YES as I do not
want any analysis from it yet until I have finished
providing it with the full article. The respond can be
seen in the screenshot below.

159
If you do not frame the prompt to specify that
ChatGPT need to wait until we have finished
providing the articles manually, it will straight away
comment on the pasted section. That would make the
analysis half-baked and not accurate.

I then proceed with the second part of the article

using this prompt which highlighted that it is a
continuation of the first part of the article.

[I will provide you with continuation of that article.

Please read and understand the article. After that, I
will ask you questions regarding the articles. I will

160
paste the articles in batches due to word limit
restrictions. Please only respond with a YES after I
have pasted the articles, and then I will proceed with
the processes until it is done. The continued article :
(paste the next part of the article here)]

Using this prompt ensures that ChatGPT still wait for

the completion of us pasting the article manually
before it engages with the task that we will provide as
can be seen in the screenshot below.

161
I then continue with the last part of the article using a
simpler prompt:

162
However, ChatGPT responded instantly after the
prompt as I did not specify that I want it to wait for my
further instruction after I have pasted the last part of
the article. Thus, it can be concluded that for
ChatGPT 3.5, you need to repeat the prompts as it
does not reflect so much on the previous criteria
within the last prompt.

163
We want to control ChatGPT to only respond after it
has analysed the whole article, therefore, the
limitation in the criteria provided within the prompt is
needed for every prompt. Only then can ChatGPT
respond as an informed reader of the whole article.
This ensures better understanding, clarity and
cohesiveness on ChatGPT's responses.

Once ChatGPT have read, understood and

comprehended the article, I can now start asking
questions regarding the article. This can be seen in
the next screenshot. The prompt that I use is:

[Can you please give me a summary of the article?]

164
As can be seen above, ChatGPT then can give a
summary on the article that I have provided.
However, bear in mind that we must still perform due
diligence by checking whether the summary is
correct. This is to ensure the reliability and the validity
of the information given by ChatGPT.

Nonetheless it does helps to speed up the process of

reading by giving us an idea of the article. If you want
to use it for your literature review, then you need to
still carefully read the article and extract the
information yourself.

There are many questions that you can ask ChatGPT

once it has read the article you have provided. This is
almost the same as you are consulting a friend whom
had read an article. I proceed my chat with ChatGPT
using the next prompt:

[Can you summarize the method used for this

article?]

The response can be seen in the next screenshot.

165
I then proceed with asking for the summary of the
findings for the article. However, I have asked it to be
presented in a table form for easier reading and
access rather than the conventional linear form. The
prompt used is:

[please summarize the findings of this article in table

form]

166
These prompts and questions to ChatGPT regarding
the article that I have uploaded should serve as initial
ideas to you on what can you ask ChatGPT to do with
the information you have provided it with. Another
useful thing to ask can be seen in the next prompt
detailed in the next screenshot. The prompt:

[can you summarize the past studies cited in this

article. Do a table categorizing them according to
author name, quotations cited and etc]

167
The continuation of this response can be seen in the
next screenshot. However, it can be seen that
ChatGPT can only capture 7 authors cited in the
article that I have provided. It also added a column on
findings, which I think is interesting as I only prompt it
in the category of [and others]. Further prompting
needs to be done maybe by including the phrase [all
of the authors cited] or any other more phrases
asking for more detailed response from ChatGPT.
This is when your creativity comes into play.

168
Next, I ask ChatGPT to highlight the information from
specific parts of the article such as the
recommendations for future study using this prompt:

[please highlight the recommendation for future

research and the gaps that this study fills]

169
In the previous prompt, I asked for two things. The
recommendations for future study and the gap that
the provided research fills. ChatGPT analyses the
article and segment its response into two parts; the
recommendation for future studies as can be seen
above and the gaps that this study fills as can be
seen below.

170
Next, I took my questions up a notch by not just
asking ChatGPT on the information within the article.
I asked ChatGPT to suggest future research based
on the research that I have provided which is the
article we have been discussing. I used this prompt:

[Can you suggest future research based on this

article]

171
These are ideas on what to ask for when you have
pasted an article onto ChatGPT. You can ask and do
so much more depending on your creativity in using
ChatGPT. Imagine what would you ask to a friend
who is an expert in an article. ChatGPT can give you
almost the same response as asking your friend.

You can also paste more than one article and then
the discussion can be more extensive, critical and
analytical. However, when doing it manually,
remember to do this process one article at a time to

172
ensure accuracy of the information. This is also
because there are not much option when you are
using ChatGPT 3.5. It has to be done one part at a
time, and one article at a time, before you could have
amassed enough information to synthesize, compare,
or contrast between studies.

The video tutorial of literature comprehension using

ChatGPT 3.5 can be accessed using the link below.

Literature comprehension using ChatGPT 3.5

Next, I will demonstrate the other three methods of AI

assisted literature comprehension using plugins,
advanced data analysis and Google Chrome
extensions from ChatGPT 4. The processes of
providing the literatures will be different, in the sense
that it will be much easier and less time consuming.
However, what you can do with the article is almost
the same. What differs is the quality of the response
and what can be done with it.

173
Literature Comprehension Using ChatGPT 4
(Plugins)

ChatGPT Plus, which offers ChatGPT 4, comes with

a paid subscription of USD 20. Personally, I think that
this subscription is very much worth it as there are so
much that can be done. This is not just limited to
academic writing, it also includes other tasks as well.
In terms of academic writing, the use of ChatGPT
Plus is not just limited on the literature review section,
but can be utilised for all parts of research process
and research writing. However, staying true to the
title of this eBook, I will focus the use of ChatGPT
Plus for literature review, particularly for literature
comprehension as discussed in this section.

We first began with identifying the plugins that can

aid the process of literature comprehension. As these
academic articles are usually in PDF format, utilising
plugins that can manipulate these PDFs is suitable
for this purpose. I will start with the search process
for these PDF related plugins. For the processes of
activating the plugins, you can refer to previous
section where I have explained this in detail. The
search for these plugins can be seen in the following
screenshots:

174
For this demonstration purpose, I have chosen Aaron
Chat PDF as the plugin. There is no particular reason
for choosing this plugin. You can choose any plugins
that can do the task based on your preference.

175
When using plugins to analyse PDF, you will need to
upload the PDF's URL (link) onto the message box in
ChatGPT. The plugins then will be able to analyse
the PDF as can be seen in the following screenshots.

I then use the simple prompt:

[Give me the summary of the article]

176
You can chat with the PDF and ask any questions
that you want. This aids in faster information retrieval.
You also have access to GPT4 while chatting with
these PDF, which is an advantage when analysing
documents as it can provide a more quality data
retrieval and insights from the academic article in the
PDF format as can be seen in the following
screenshot.

177
The video tutorial of literature comprehension using
ChatGPT 4 (Plugins) can be accessed using the link
below.

Literature comprehension using ChatGPT 4 (Plugins)

Next, I would like to demonstrate another option of

reading and comprehending PDF files using
Advanced Data Analysis, embedded within ChatGPT
4. This is a faster method, however it is a little bit
limited in terms of the response it can give.

Literature Comprehension Using ChatGPT 4

(Advanced Data Analysis)

Utilising Advance Data Analysis already embedded in

ChatGPT 4 is a much simpler and faster way of
analysing PDF text such as academic articles which
makes a large bulk of our literatures. This process is
faster compared to the two previous method
explained before as you can directly upload
documents, in our case, PDFs, using the upload icon
available at the message section as illustrated in the
following screenshot.

178
You can upload up to 10 documents at a time to
ChatGPT 4. However, based on my experience,
uploading fewer number of documents increase the
quality of the response from ChatGPT in analysing
these documents. It is advisable to upload in small
batches to ensure a smooth flowing chat session and
quality output. I have demonstrated this in the
following screenshot by uploading three (3) PDF files.
You can also straight away prompt on the task that
you want ChatGPT to do with the uploaded PDF as
can be seen in my prompt:

179
[Please analyse these articles and compare and
contrast the findings in a table format]

This surely saves time as I did four task at one time.

One, I have uploaded three PDFs. Two, I ask GPT4
to analyse the articles. Three, compare and contrast
the findings of these articles and four, present them in
a table.

Comparing this with the manual PDF analysis using

3.5 is a huge step up in terms of effort and time
consumed.

180
Once we have prompted ChatGPT 4, it automatically
used Advanced Data Analysis to analyse the articles
as can be seen in the following screenshot. Notice
how in the prompt I did not specify ChatGPT to use
Advanced Data Analysis to analyse these articles?
ChatGPT detects that it needs to use it automatically.
This saves time in terms of prompting.

181
One feature that stands out when using Advanced
Data Analysis is the ability to analyse multiple
documents at one time. This is very useful in
comparing and contrasting between different studies.
The formation of a table such as the ones displayed
in the following screenshot is a valuable output from
the analysis of these articles. You can replicate the
process and ask ChatGPT to update the table.
Furthermore, you can even ask ChatGPT to produce
the table in an editable MS Word format, which saves
more time. I have demonstrated this creative
instruction of asking for an MS Word file in the
previous section on developing literature matrix.

182
The prompts given to ChatGPT as a guide to
comprehend these articles are just the tip of the
iceberg. You can do so much more based on your
needs and creativity. You can even ask content
related questions to ChatGPT based on these
articles.

However, please be informed that the information

generated by ChatGPT in analysing these articles
should always be checked for accuracy as there is no
guarantee that the information are accurate. Another
important criteria of analysing articles in the ability to
grasp the content of the article as a whole, and come
up with the gist of the article.

Even though AIs, particularly ChatGPT in this case,

can do that, it cannot beat human comprehension
over academic articles. Careful reading and
comprehension of academic article is more accurate
and the information extracted is more valuable.
Therefore, this AI aided reading and comprehension
of documents can be used to serve as providing initial
drafts of literature analysis, which should be
complimented by the conventional reading and
comprehension of text. You must read your
literatures. Do not fully automate your literature
reading and comprehension!
183
The video tutorial of literature comprehension using
ChatGPT 4 (Advance Data Analysis) can be
accessed using the link below.

Literature comprehension using ChatGPT 4 (Advance

Data Analysis)

Literature Comprehension Using ChatGPT 4

(Google Chrome Extension)

Another option for literature reading and

comprehension using ChatGPT is the utilisation of its
Google Chrome extension. ChatGPT allows its
extension to be embedded in Google Chrome. This
means that everything that is displayed on your
browser can be analysed by ChatGPT. You can
summarise, paraphrase, translate and do many other
task. Your creativity is the only limitation in using this
extension.

This is indeed a valuable feature that can be used for

literature comprehension. Once you have opened an
academic article on your browser, you can access
ChatGPT 4 using the extension buttons that are
highlighted in the following screenshot. You need to
install the extension into your Google Chrome before
you can do this. The installation process is very
simple.
184
In the example illustrated in the following screenshot,
I have opted for ChatGPT to summarise the article.
You can do many other task as I have stated before.

185
For demonstration purpose, I have further prompted
ChatGPT 4 (via the extension in Google Chrome) to
summarise the methodology of this study used as an
example.

186
As the article used as the example is a review article,
therefore, the methodology might be different from
what usually would be extracted from an academic
paper. This extension can respond in the same
quality and length that ChatGPT 4 can do when you
access it directly.

For demonstration purpose, I then prompted

ChatGPT 4 in extension of Google Chrome to
arrange the information in table form.

187
As this is a review article, the summary of findings
contains findings from other studies. ChatGPT also
extracted the authors for each findings reviewed and
a link was created for the article of each of the author.

188
The use of ChatGPT 4 as a Google extension
promotes efficiency and saves a lot of time. You can
speed read and speed analyse an article before you
decide whether to download the article. You can save
time in terms of determining the relevancy of the
articles that you have searched for before deciding on
further actions with it.

For some, the translation feature is also a valuable

feature to assists better comprehension. The ability to
paraphrase, summarise, shorten and lengthen text in
real time of reading aids a much faster reading and
text extraction process.

The video tutorial of literature comprehension using

ChatGPT 4 (Google Chrome Extension) can be
accessed using the link below.

Literature comprehension using ChatGPT 4 (Google

Chrome Extension)

These are great features for initial reading of

literatures before we deeply engage with the
literatures that are relevant for our research project.

189
Next we will explore the use of ChatPDF for literature
reading and comprehension. Please do not confuse
ChatPDF with ChatGPT. ChatPDF is a different
service altogether from ChatGPT. It is not a plugin or
an extension of ChatGPT. It is its own site and is
dedicated solely on interacting with PDF.

Literature Comprehension Using ChatPDF

You might be wondering, since ChatGPT can already

be used to analyse literatures, then why do I still
introduce ChatPDF? The answer is that, based on my
experience, I believe that this is the most accurate
and useful PDF extractor in the market. This is my
preference. I understand that there are many other
useful PDF extractor that you may prefer. You can
use any PDF extractor that you like. As long as you
are comfortable with it and it gets the job done, then
that is the most suitable AI for you.

ChatPDF offers free usage however this is a bit

limited. The free account is actually quite good, so if
that is sufficient for you then you can just opt for the
free account. The paid version (at around USD 5 per
month) offers analysis for 2000 pages per document
at the size of 32MB per documents.

190
This means that you can analyse books using
ChatPDF. The screenshot below portrays what to
expect on your first contact with ChatPDF.

To start interacting with PDF, You can drag and drop

your PDF file in the designated area.

Within seconds it had already analysed the PDF

given and even gave sample questions to ask. These
questions are quality suggestions based on the
content of the PDF and not just simple questions and
task such as summary of the article.

191
It can list the key findings from the academic paper
uploaded, equipped with the pages where the
information can be found. You can click on the page
number and you will be directed to the page, and the
paragraph where the information were extracted from
will be highlighted for you.

192
The same thing can be seen when I asked on the
content of the article. Reference towards the page
number of the information is a useful feature as it
ensures reliability of the ChatPDF's response. We
can easily check whether the information given is
accurate, simply by clicking on the page number icon.

193
It can even analyse books within seconds. The book
illustrated in the following screenshot is around 600
pages. You can even ask questions where it
responds by analysing the content of the pdf, and not
just reiterating the content of the pdf to you as can be
seen in the following screenshot where I used this
prompt:

[I am doing content analysis on hate comments on

Facebook where I will collect data from Facebook
group that were created to discuss on these hate
comments. I will ask members from the group to

194
provide me with hate comments that they have
received. what kind of data collection is this data
collection?]

ChatPDF were able to comprehend the question, sift

through 600 pages from the book and then suggest
the method of data collection that I was looking for.

195
Of course I will further read and validate this
suggestion, hence why I ask for further elaboration to
be given complete with the page number reference.
However, I have saved a lot of time to locate for the
information as my reading is now aided with
ChatPDF.

The video tutorial of literature comprehension using

ChatPDF can be accessed using the link below.

Literature comprehension using ChatPDF

196
Another valuable feature of ChatPDF is that you can
also effectively chat with multiple PDF by combining
them in a folder. You will need to first create a folder.
Then you need to upload the pdfs one by one. After
that you can just drag and drop the pdfs from the
toolbar onto the folder that you have created as can
be seen in the following screenshot.

197
For demonstration purpose, I created a folder with 6
academic articles on the use of citations in academic
writing. I named the folder 'citation'. You can add
more articles in your folder. Then, comparisons
between these pdf can be made on the content that
you want. You can compare and contrast between
these pdfs. The compared items and information will
be highlighted using the numbers as can be seen in
the following screenshot. When you hover your
cursor on the numbers, you will see information such
as the name of the article and the page number these
information were extracted from.

198
With the advanced capabilities of AI tools like
ChatPDF, you are not limited to merely extracting
information from PDF documents. One of the
remarkable features of ChatPDF is its ability to
provide informed suggestions and insights. This is
possible because ChatPDF does not just passively
scan the documents. It actively analyses and
understands the content within these PDFs. As a
result, when you pose questions or seek guidance,
the AI can leverage its accumulated knowledge from
the analysed texts to offer suggestions that are
contextually relevant and insightful. This feature
transforms the way we interact with literature,
allowing for a deeper and more holistic
comprehension. It is like having an intelligent
assistant that not only comprehends the material but
also contributes meaningfully to your understanding
of it. The following screenshot illustrates this.

199
The video tutorial of literature comprehension using
ChatPDF (multiple documents) can be accessed
using the link below.

Literature comprehension using ChatPDF (multiple

documents)

200
In this chapter I have shared with you the exciting
ways in which ChatGPT can enhance your
understanding of literature. Let me once again
reiterate the four innovative methods. Firstly, I have
discussed the manual use of ChatGPT 3.5, where
you can engage in interactive sessions to
comprehend deeper into literary themes and
concepts. Secondly, I have explored the use of
ChatGPT 4 with plugins, which broadens the AI's
capabilities, allowing for a more thorough analysis.

Thirdly, I have experimented with ChatGPT 4's

advanced data analysis feature, offering you a richer
comprehension of complex texts. Lastly, I introduce
the use of a Google Chrome extension with ChatGPT
4, providing a convenient and integrated tool for
literary insights right in your browser. Additionally, I
explained how ChatPDF can be a very useful tool in
literature comprehension by extracting and analysing
content from multiple PDF documents, offering
structured and detailed insights.

I want to highlight the significant advantages of using

AI in literature comprehension. Tools like ChatGPT
and ChatPDF can save you a lot of time and effort.
They are particularly useful for creating initial drafts

201
and aiding in reading and comprehension. These AI
tools provide quick access to information and insights
that might otherwise take a lot of time to gather. This
efficiency is of great help, especially for students and
researchers who often operate under tight schedules.
Nonetheless, it is important to remember that these
tools are best suited for gaining a preliminary
understanding and assisting in the early stages of
literature review and analysis.

However, and this is crucial, I must stress the

irreplaceable value of conventional methods of
literature comprehension. While AI tools offer
significant assistance, they cannot replace the depth
and accuracy of human understanding. This is
particularly true for texts that are highly relevant to
your field of study. Conventional methods of reading
and analysing literature provide insights and a
personal connection that AI simply cannot replicate.

In conclusion, I want to remind you that while AI can

be a helpful tool in your literature comprehension
journey, it should be used as a supplement to, rather
than a replacement for, the rich understanding that
comes from your personal engagement with literary
works.

202
 Part IV
Literature Review
Writing
 W R I T I N G
S K I L L S F O R
L I T E R A T U R E
R E V I E W
In the process of conducting research, whether for
academic articles or theses, the literature review
takes on various forms, each defined by its objective
and approach. These forms are outlined in a simpler
manner as follows:

1. Narrative or Conventional Literature Review

This type involves a thorough and unbiased review of

existing knowledge on a topic. It summarizes the
main findings from relevant research and identifies
areas where researchers agree or disagree.

2. Systematic Literature Review

This kind of review is detailed and aims to be as

unbiased as possible. It involves carefully planning
and searching for all the studies on a particular topic,
mainly used in fields like health and social sciences.

3. Meta-Analysis

A specific type of systematic review, meta-analysis

uses statistical methods to combine results from
different studies. It is helpful for bringing together
data from smaller studies to make more reliable
conclusions.

205
4. Scoping Review

This review is often the first step in a research

project. It helps to map out the key ideas and types of
evidence in a research area, especially when the
topic is new to the researcher.

5. Critical Literature Review

This approach involves critically analysing and

combining different pieces of literature. It helps to
create new ways of looking at a topic.

6. Integrative Literature Review

This type combines literature on a topic in a way that

creates new ideas and viewpoints. It looks at all
studies that deal with similar or the same questions.

7. Theoretical Literature Review

Focused on gathering and analysing theories related

to a specific issue or concept. It is useful for
understanding what theories exist, how they relate to
each other, and what has not been studied yet.

206
8. Methodological Literature Review

This review looks at the methods used in previous

research and their results. It is useful for figuring out
how to approach a research question and for finding
new ways to interpret data.

Each type of literature review serves a different

purpose and can be used in various parts of a
research project. The choice of which type to use
depends on the research question, the academic
field, and the researcher's preferred method.

207
A I A S S I S T E D
L I T E R A T U R E
R E V I E W
D R A F T
The integration of Artificial Intelligence (AI) into the
process of creating literature review drafts marks a
significant advancement in academic research
methodologies. AI-assisted tools are redefining the
way literature reviews are conducted by offering
unparalleled efficiency and analytical depth. Utilising
natural language processing (NLP) and machine
learning (ML), these tools can swiftly parse through
extensive databases of academic papers, discerning
relevant studies and extracting key findings. This
capability allows researchers to produce draft reviews
that are not only comprehensive but also meticulously
tailored to their specific area of inquiry. The result is a
draft that captures the essence of existing
knowledge, setting a solid foundation for further
research.

Moreover, AI assistance in drafting literature reviews

ensures a level of consistency and
comprehensiveness that might be challenging to
achieve manually. AI algorithms are capable of
identifying and synthesizing research trends,
discrepancies, and gaps in the literature with minimal
bias, which is crucial for the integrity of a literature
review. As AI technology continues to evolve, these
tools are becoming more adept at understanding

209
complex academic discourse, thereby enhancing the
researcher's ability to draft reviews that are both
informative and critical. The draft produced with AI
assistance is not the final product but a sophisticated
starting point that provides a clear direction for the
subsequent research journey, ensuring that the final
literature review is of the highest possible quality.

In this eBook, we experiment with a practical

framework for leveraging two distinct AI platforms that
is ChatGPT and Scite in developing a robust
literature review drafts. ChatGPT, with its
conversational AI capabilities, excels at
understanding and generating human-like text, which
can be instrumental in articulating complex ideas,
summarizing research findings, and crafting a
narrative that interweaves various strands of research
into a cohesive story.

Scite, on the other hand, provides an innovative

approach to assessing the veracity and impact of
scientific articles by using smart citation metrics,
which aids in discerning the most influential and
relevant studies.

210
It is hoped that the incorporation of both these AI
technologies, can showcase a harmonized process
where ChatGPT's generative strengths complement
Scite's analytical precision. To elucidate this synergy,
I will include more demonstrative screenshots,
guiding you through the step-by-step process of
constructing a literature review draft that is both
comprehensive and critically analysed, thus offering a
tangible and visually enriched learning experience.

Literature Outline Drafts Using ChatGPT

Allow me to start our journey for literature outline

drafts with ChatGPT using a simple prompt:

[I am conducting a study on the impact of digital

marketing using social media on Micro Small Medium
Enterprise (MSME) in Malaysia. Now I want to write
my literature review. Can you suggest an outline draft
for my literature review?]

211
212
This is the outline generated by ChatGPT as a
suggestion for my prompt. I then further prompt
ChatGPT using this prompt:

[I am using the funnel strategy for my literature review

section. Can you suggest an outline based on the
strategy?]

213
I can further prompt ChatGPT to change the outline
of this literature review draft according to my
preference as can be seen in the next prompt:

[Thank you. I see that you started the outline with

Digital Marketing. I want to start my literature review
with MSME. Can you provide an outline suggestion
for this?]

214
There is yet another spelling mistake in the prompt as
I have prompted SMSE whereby it should be MSME.
However, I just let it be as ChatGPT still gives the
correct output, and even corrected this in its reflection
of the prompt. This just proves that to an extent,
ChatGPT is an intelligent and capable AI.

The continuation of ChatGPT's suggested outline can

be seen in the following screenshot.

215
As can be seen in the last two previous screenshots,
ChatGPT has changed the structure of my literature
review outline draft. I can change this as many times
as I like, in any way or structure that I prefer and
deemed suitable as can be seen in the next prompt:

[Thank you. Now I want to start my literature review

with Social Media as my research is actually in the
domain of new media literacy. Please suggest an
outline for that]

216
217
Please note that these alterations to the literature
review outline drafts serve as an example in
showcasing the capability of ChatGPT in doing these
task. It does not serve the purpose of showing you
how your literature review outline should be
structured.

Now that I already have an outline that I am satisfied

with (let's just assume that I am satisfied with this
outline), I can further venture into the sub sections of
the literature review outline, that is sub-section 3
(refer to the previous screenshot) as demonstrated in
the following screenshot using this prompt:

[Good job. Now can you help me to develop an

outline for sub section 3: Social Media in Business
and Marketing]

218
The continuation of the respond is illustrated in the
following screenshot.

Literature review outline draft using ChatGPT.

219
The video tutorial of writing literature review outline
drafts using ChatGPT can be accessed using the
link below.

Moving deeper in the conversation, I then can

prompt ChatGPT to write me a suggested draft for
this subsection as illustrated in the following
screenshot using this prompt:

[Thank you. Now can you help me to write a

suggested draft for this subsection]

220
The continuation of the response can be seen in the
following screenshot.

221
I then can go into more detail as illustrated in the
following screenshot where I asked ChatGPT to
further eleborate the subsection 3.2 which is The
Role of Social Media in Modern Marketing using this
prompt:

[Can you help me to elaborate 3.2]

The continuation of the response is illustrated in the

following screenshot.

222
I then experimented on this draft using Litmaps to see
whether it can find me relevant studies related to this
particular draft. I use this prompt:

[Using litmaps, can you edit this writing by suggesting

relevant, current and high impact literatures
embedded as in text citation using the APA citation
format: (I pasted the text here)]

The following screenshot illustrates the result of this

prompt.

223
Litmaps then suggested a few literatures that it
deemed is related to my literature draft as can be
seen in the following screenshot.

224
I then ask a question which answer might be a shock
to you:

[Do you have the ability to analyse these articles?]

The answer, as you can see in the following

screenshot, is that Litmaps actually only read the
metadata and the abstracts available in the Litmaps
database only. Therefore, these suggestion of article
are just merely a superficially intelligent suggestion
which could not be accounted on its reliability and
validity.

That is why, we can only use ChatGPT for producing

drafts (in this case an extensive literature outline
draft) and cannot sub-contract our writing to
ChatGPT. This reason transcend ethical issues which
can be tolerated and manipulated according to our
belief system and value. This concerns reliability and
validity of our literature analysis which could not be
tolerated in any academic setting and environment.

225
226
That is why, we can only use ChatGPT for producing
drafts (in this case an extensive literature outline
draft) and cannot sub-contract our writing to
ChatGPT. This reason transcend ethical issues which
can be tolerated and manipulated according to our
belief system and value. This concerns reliability and
validity of our literature analysis which could not be
tolerated in any academic setting and environment.
However, to satisfy our curiosity, I still prompt
ChatGPT to integrate the studies it has suggested,
into the draft produced, as can be seen in the
following screenshot using this prompt:

[Edit the writing using the study that you have

provided using Litmaps and integrate them in the
writing using in text citation APA format]

As illustrated in the following screenshot, ChatGPT

can do that. Superficially, the writing appeared to hold
some credibility as the writing were embedded with
citations in the APA format. However, there is no

227
other way to determine whether these citations were
placed correctly in terms of the reference to the
content it was cited for other than for us to carefully
read the article.

However, we can now aid our reading as I have

demonstrated earlier in the chapter on AI assisted
literature comprehension. That should save us a
considerable amount of time. Nonetheless, reading
still need to take place.

We need to always bear in mind that ChatGPT was

never designed to be able to do our research for us. I
acknowledge that with creativity, something quite
similar to that can be achieved. However, we still
need to do due diligence of providing ChatGPT with
the relevant literatures.

It can do a better job if we can provide it with the

summarised findings of literatures. Then, the output
by ChatGPT can be valuable and reliable. This is
where you train your GPT with the knowledge
necessary and suitable for the task that you expect it
to do.

228
However, what we have now is an extensive draft for
the outline of our literature review. This should be
more than enough to help you kick start your writing.

The video tutorial of writing literature review drafts

using ChatGPT can be accessed using the link
below.

Writing literature review drafts using ChatGPT

229
A more quality response can be obtained if we
integrate the various methods, strategies and
techniques that have been demonstrated in previous
chapters such as literature search, management and
comprehension before we start producing drafts for
our literature review.

Next we will explore another powerful AI for research

that is Scite. It is able to write with citations and there
are so much more features that can be explored..

230
Literature Review Drafts Using Scite

Upon the completion of a detailed outline, assisted by

ChatGPT, our focus shifts to Scite, an essential tool
designed to enhance our literature review. Scite
extends beyond mere citation counting; it offers an
analytical perspective by categorizing each citation as
supportive, contradictory, or neutral in relation to the
research under scrutiny. This level of discernment is
vital, as it affords a comprehensive insight into the
research's standing within the scholarly dialogue.

With an AI-assisted outline in hand, we utilize Scite to

bring depth and complexity to our literature review.
Scite's novel approach to citation analysis provides a
comprehensive understanding of not just the quantity,
but the significance of scholarly discussions. This
chapter will demonstrate to you on how to integrate
Scite's functionalities into your writing process. This
can ensure that your literature review will not only be
exhaustive but also critically enriched by the diverse
viewpoints within the academic sphere.

The video tutorial of the key features of Scite can be

accessed using the link below.

Key features of Scite

231
Literature Review Drafts Using Scite (Assistant)

I am going to start by using one of Scite's features

(Assistant) to suggest a background of the literature
draft for the previous outline given by ChatGPT using
this prompt (which includes the pasted outline
component as previously suggested by ChatGPT:

[Please help me to write a draft for this outline:

Background.
Overview of the MSME sector in Malaysia: Size,
significance, and contribution to the Malaysian
economy.
Brief history and evolution of digital marketing.
The role of social media in modern marketing
strategies.]

How Scite process the prompt using its Assistant

feature can be seen in the following screenshot.

232
As can be seen in the screenshot above, Scite
composed its response based on 25 result. This does
not mean all the 25 searches were included in the
response. Scite chooses the most relevant literature
to our prompt and construct response based on that.

The following screenshot illustrates Scite's final

response for this prompt. It was able to integrate past
studies into its response using in-text citation in the
APA format.

233
Scite integrates past studies, and cite them in its
response without being prompted to do so. As it is
designed for this purpose, we do not have to request
for citations, as what we need to do with ChatGPT.
This is the advantage of using a specialised AI that is
specific for a specialised task such as Scite for
literature review writing related task and ChatPDF for
PDF information extraction.

There are 7 past studies cited in this response. In my

opinion, the draft that Scite produced is actually quite
good, in terms of its academic writing. It suits a
response for a background of the literature based on
the topics prompted.

You can view the keywords used and publications

consulted by Scite in producing its response by
clicking on the area that I have highlighted in the
following screenshot.

234
In the following screenshot, I would like to
demonstrate how we can determine whether the
citations given by Scite are accurate. I demonstrate
this using the study that was cited in the response,
that is the study by Mason et al. (2021).

235
As can be seen above, the detail for the citation is
given in the references area beside the response. It
contains information such as the exact quotation of
the citation, the section where the quotation is taken,
the title of the paper, the name of the authors, the
journal of the article, the number of publication citing
this work and some other detailed information such
as the number of supporting, mentioning and
contrasting citation statements for this work. You can
also cite using the cite feature as can be seen in the
screenshot above. For further demonstration of these
features you can access to my video explanation via
the link below.

236
I then click on the title of the paper and these
information were presented in a more detailed and
user friendly manner. I wanted to have access to this
academic article, therefore, I click 'View full text' as
can be seen in the following screenshot.

After clicking the 'View full text', I was directed to the

source of the academic article.

237
As this is an open access article, I was able to
download it. However, this does not mean that if it is
not an open access article, you will not be able to
download it. As we are directed to the publisher's site,
you can access using your institution's subscription
and download the article. The following screenshot
shows the downloaded academic article.

238
In order to ensure that the citation statement
produced by Scite for this article is correct, I need to
read the paper. Thus, I integrate AI in the process of
targeted reading by uploading this article into
ChatPDF as can be seen in the following screenshot.
The process of uploading articles into ChatPDF have
been explained in the previous chapter in Part III:
Literature Comprehension within the AI assisted
literature comprehension chapter.

239
The video tutorial on writing a literature review drafts
using Scite (Assistant) can be accessed using the link
below.

Writing a literature review drafts using Scite

(Assistant)

I then prompted ChatPDF with the question:

[Is this statement stated in this study? :social media

marketing has gained importance after the COVID-19
pandemic, emphasizing the adaptability and
resilience of marketing strategies in response to
external disruptions.]

240
This is the statement that this article was cited for in
Scite's response when I prompted it for the
background of literature from the outline generated by
ChatGPT.

As can be seen in the screenshot above. Scite

confirmed that the statement is not explicitly
mentioned in the article however the article does
emphasise the growing importance of social media
since the Covid-19 pandemic. The page number
where the statement can be found was also given.

241
When I click the number of the page that was
provided by ChatPDF, I was directed to the page, and
to the paragraph where the statement can be found.
It can be found in the abstract. I know that the title of
the article already reflect the statement, however, we
cannot just cite based on the title. Even if it is
mentioned in the abstract, I am still not satisfied and
would like to be very sure about this.

Thus, I further prompted ChatPDF:

[list the key findings of this study that is related to:

social media gaining importance after Covid-19]

The screenshot below portrays ChatPDF's response.

242
There are several quotations related to my query.
ChatPDF provided these information equipped with
the page number and paragraph(s) where this
information can be found as can be seen in the
screenshot above.

The following screenshot shows the exact location

where this information was taken.

Now that we are sure that the statement actually exist

in the article, we can now use it in our literature
review. Sure, it takes a lot of task, however I believe
that you will agree with me that it has already sped

243
up the process of literature outline, drafts, search and
reading.

You can now get ideas on what to write, how to

structure it, how to write drafts on it and what to cite
faster than before you integrate AI in your literature
review process.

I understand that the ideal way would be to read

many papers, summarise and synthetize them
manually before we produce the outline and drafts of
our literature review. We should do all these
processes as it aids our understanding in order for us
to write a critical literature review. However, these
can be aided by AI as I have demonstrated
throughout the book within the literature search,
management, reading and writing process.

244
Literature Review Drafts Using Scite (Custom
Dashboard)

Another feature that is very useful in Scite is the

ability to use the Custom Dashboard feature. You will
need to select custom dashboard from the list of
product available by Scite as can be seen in the
following screenshot.

After you have selected Custom Dashboard, you will

be directed to the Custom Dashboard interface. You
must create your own dashboard using the button
provided as can be seen in the following screenshot.

245
You can import citations from Zotero or Mendeley
library as your custom dashboard.

246
You can then name and provide description for your
custom dashboard.

After that, you will need to select the literature that

you want to include in your Custom Dashboard.

247
For demonstration purpose, I have selected a folder
from my Mendeley to be used. The folder contains
academic articles related to how article authors cite in
their academic articles. You can import up to 30
academic articles per custom dashboard.

248
Once your dashboard was created, Scite will analyse
the academic articles and a report will be provided to
you as illustrated in the following screenshot.

249
You can now access to your custom dashboard using
the shortcut (/) on your keyboard.

250
Scite then will ask you to specify which dashboard
you would like to use for the task. You can choose
the from the dashboards that you have created and
add the dashboard onto the assistant.

Once added, the dashboard will be integrated in your

Assistant as can be seen in the following screenshot.

251
I then prompt Scite's Assistant to write a draft using
this prompt:

[trends of citation in academic articles]

Scite will then write using the articles available in the

dashboard only, hence the name Custom Dashboard.
This ensures control over the references used in
writing the draft. As these papers were chosen by us,
we can ensure the quality, relevancy and currentness
of the literatures used in our drafts.

252
The use of the Assistant feature on Scite for
generating literature review drafts offers a significant
advantage in terms of efficiency and
comprehensiveness, providing a robust starting point
for researchers and writers. This innovative tool
leverages the power of technology to compile and
organize vast amounts of academic literature,
streamlining the initial stages of the literature review
process. It assists in identifying key studies, trends,
and gaps in the existing research, thereby saving
valuable time and effort that can be redirected
towards deeper analysis. The Assistant's ability to
quickly gather and present relevant information from
a broad spectrum of sources is particularly beneficial
in today’s fast-paced academic environment, where
staying abreast of current research is both
challenging and crucial.

The video tutorial on writing literature review drafts

using Scite (Custom Dashboard) and ChatPDF can
be accessed using the link below.

Writing literature review drafts using Scite (Custom

Dashboard)

253
However, it is imperative to approach these
generated drafts with caution. While the Assistant
feature is a powerful tool for collation and
summarization, the task of critically analysing,
synthesizing, and contextualizing research findings
remains a distinctly human endeavour. A high-quality
literature review goes beyond mere aggregation of
sources; it requires a careful and thoughtful
interpretation of research in relation to specific study
objectives, theoretical frameworks, and
methodological considerations.

This level of analysis ensures the review's relevance,

coherence, and contribution to the field. Therefore,
writers should use the drafts as a foundation to build
upon, employing their expertise, critical thinking skills,
and academic rigor to craft a literature review that is
not only informative but also reflective of a deep and
meaningful engagement with the subject matter. This
approach ensures that the final literature review
upholds the standards of academic excellence and
integrity, contributing genuinely to the advancement
of knowledge in the field.

254
You can integrate ChatGPT, Scite and ChatPDF for
writing literature review drafts.The integration of these
three AI, in my opinion can account for the flexibility,
accuracy and effectiveness of AI usage in aiding
literature review writing.

255
 Utilising
Custom GPTs
For Literature
Review Tasks
D E V E L O P I N G
C U S T O M G P T
F O R
L I T E R A T U R E
R E V I E W
T A S K S
What is Custom GPT?

Custom GPTs, are specialized versions of the AI

language models designed to cater to specific tasks
or domains. Unlike generic AI models, these custom
GPTs are tailored to serve particular functions, such
as academic writing, legal consulting, or medical
information processing. Users can personalize these
models by specifying the tasks they need assistance
with, and the AI adapts its responses accordingly.
This customization allows for more focused and
relevant outputs, enhancing the efficiency and
effectiveness of the tasks at hand. You can have
access to this feature when you subscribe to
ChatGPT Plus.

In the context of literature review writing, a custom

GPT can be immensely helpful. It can be
programmed to understand and adhere to academic
standards, ensuring that the literature review is
thorough, well-structured, and aligns with scholarly
conventions. The model can assist in identifying
central themes and key research areas in existing
literature, suggesting relevant sources, and offering
insights into how these can be integrated coherently
into the review. It can also provide tailored advice on

258
organizing the review, maintaining a critical and
analytical approach, and ensuring proper citation and
referencing. This personalized assistance not only
streamlines the literature review process but also
elevates the quality of the final output, making it a
valuable asset for students and researchers in
postgraduate studies.

Developing Custom GPT

The processes of developing your own custom gpt, in

this case for the purpose of literature review, are
illustrated using the screenshots below.

259
First, you will need to click the option 'Explore GPTs'
as illustrated in the screenshot above. This will give
you access to GPT store and the 'My GPT's' option
illustrated in the following screenshot.

Then, you need to click the button 'create' in order to

develop you own GPT that can be tailored to your
specific needs.

260
The following screenshot displays the interface for
GPT Builder. This is a room where you build your
GPT. I then use this prompt as the initial instruction
for this demonstration purpose:

[Hi. I would like to create a gpt that assist me in my

literature review writing for academic research article]

ChatGPT then process the instruction and created

the base of the GPT. The base can be seen at the
area on the right hand corner of this interface. It is
referred to as the 'playground' for custom GPT. This
is where you test the custom GPT before you launch
it. It then suggest a name for my GPT. I do not like
the name, hence I further prompted it to change the
name of my GPT to be 'Literature Review Assistant'
as illustrated in the next screenshot.
261
The next screenshot shows the process of ChatGPT
updating my custom GPT based on my instruction. It
does not only update the name, but now is creating
the profile picture for my GPT

262
Next, I activate the web-access option in ChatGPT
and then prompted my custom GPT to search the
Internet using this prompt:

[Please browse the Internet and learn on how to write

a good literature review]

263
As can be seen in the following screenshot, ChatGPT
then browsed the Internet in search of this
information to be updated into my custom GPT's
knowledge. This is just one option of educating our
custom GPT. We can also upload documents such as
books on literature review writing and use that as a
base knowledge for our custom GPT.

Once ChatGPT has completed its learning, I then

prompted my custom GPT for specify on the
language output that I expect it to produce. This
demonstrates the personalisation aspects that can be
achieved by using custom GPTs. I provided the
sample of language that I want my custom GPT to
emulate using this prompts:

264
[I will give you some articles that I deemed are
exemplary in terms of its literature review section.
Please use languages that emulates the language
style, tone, vocabulary and structure such as these
samples. Analyse the texts thoroughly and your
production afterwards should be equivalent (in terms
of language use) with these articles. Do not plagiarise
and do not use the content from these articles. Just
emulate the language].

I then choose the documents that I want to upload

into my custom GPT. The uploading process can be
done using the upload button readily available at the
left hand side of the chatbox, or you can just simply
drag and drop the document onto the instruction
area.

265
The next screenshot illustrates the interface once I
have successfully uploaded the document. I chose
five research academic articles to be used as
language sample for this custom GPT. The more
document you upload in your custom GPT to be used
as its knowledge, the better the result will be.
However, please note that only ten documents are
allowed to be uploaded as the knowledge for each
custom GPT.

266
ChatGPT then process these documents and update
them as the knowledge for this custom GPT as can
be seen in the following screenshot.

267
Once completed, ChatGPT will suggest you to try the
custom GPT in the playground area. This is to better
fine-tune the custom GPT to meet your expectation.
Based on the response that you get in the playground
area, you can decide on what else to train your
custom GPT with, or what prompts should you
instruct your custom GPT in order to make it better

268
Once you are satisfied with the custom GPT, you will
need to publish and save the GPT. There are three
options to choose from such as 'only me', 'anyone
with a link' or 'everyone'. The screenshot below
shows four option because I am subscribed to
ChatGPT Teams, therefore, I have the option to
publish this custom GPT to my team only. If you are
using a personal account, then you should only have
three option.

You can choose any option that you want. Once you
click 'save', your GPT will be published and you can
start using it.

269
The screenshot below shows the interface of my
custom GPT.Once published, you can start using
your custom GPT. A shortcut to your custom GPT will
be created and can be accessed at the top left corner
of your ChatGPT toolbar.

Even after you have published your custom GPT, you

can still edit it to better meet your expectation. The
key to create a good custom GPT is training it with
the correct set of data and to effectively prompt it
during the development as we are now coding the
custom GPT using language. Be creative in
developing your custom GPT and it will be a great
help for its intended purpose.

270
In the next chapter, I will demonstrate how we can
harness the custom GPTs made by others from the
GPT Store. This is an easy, and sometimes more
effective way of using ChatGPT to aid the process of
literature writing.

Creating literature review related Custom GPT.

271
H A R N E S S I N G
L I T E R A T U R E
R E V I E W
R E L A T E D
G P T S F R O M
G P T S T O R E
Custom GPT models developed by others and
available in the GPT store represent a significant
advancement in the field of specialized AI
applications. These models are crafted by experts in
various domains, ranging from academic research to
niche industry sectors. The primary advantage of
using these custom GPTs is their expertise in specific
areas. Since they are tailored by professionals with
deep knowledge and experience in particular fields,
these models offer insights and guidance that are
both nuanced and highly relevant.

The use of these custom GPTs can be particularly

beneficial for several reasons. First, they save time
and resources. Instead of building a model from
scratch, users can leverage the expertise embedded
in these pre-built models. Second, the quality of
output is often superior for specialized tasks. These
models have been fine-tuned on domain-specific
data, making them more adept at understanding and
responding to complex queries in their respective
areas. Third, they offer a level of customization and
flexibility. Users can choose a model that closely
aligns with their specific needs, ensuring more
accurate and efficient assistance.

273
In scenarios where precision and specialized
knowledge are crucial, such as academic research,
literature reviews, or technical writing, using these
specialized GPTs from the store is often preferable.
They not only enhance the accuracy of the
information provided but also ensure that the output
adheres to the standards and practices of the specific
field, thereby adding significant value to the user’s
work or research.

The GPT Store can be accessed by clicking the

'Explore GPTs' button on your ChatGPT toolbar as
shown in this screenshot.

274
You will then be directed to the GPT Store as can be
seen in the next screenshot. You can search for the
GPT that you deemed best fit your intended purpose.
These GPT were created by companies and
individuals and were published on the GPT Store.
The GPT Store is the same concept with an AppStore
by Apple and PlayStore by Android. The idea is to
pair developers with users.

275
For this demonstration purpose, I have used the term
'research' for my GPT search. I then were suggested
GPTs that can aid in research process. One of the
advantage of using GPTs that were developed by
others as opposed to building our own GPT is that
these GPTs were created by developers using coding
and APIs thus increasing their effectiveness. As
some of these popular GPTs were created by
companies who also have their web-based service
readily available, their efficiency can be greater than
our own custom GPT which were only coded using
language and does not have a web based service
attached to it.

276
However, please note that if you are good in coding
your custom GPT, then using your custom GPT can
yield better result than selecting one from the GPT
Store. A custom GPT that we build also ensures
better personalisation of GPT function. However,
choosing from the GPT Store is easier and more time
effective. Balance between the two and you can truly
harness the function of these custom GPTs.

For this purpose, I chose the GPT 'Consensus' as

can be seen in the following screenshot. Consensus
is a research specialised web service that can be
accessed on the Internet. It assist in finding relevant,
current and high impact research articles and
literatures. As a GPT, other than can be used to
search for literatures, Consensus also allows us to
draft content with accurate citations. It is a very
popular GPT in the GPT Store.

277
The following screenshot displays Consensus work
process once you have instructed it to do something.
It will redirect the request onto its web based service
and you will need to allow such interaction in order for
Consensus to be able to process your instruction.

278
To showcase what Consensus can do, I start this
demonstration with the prompt:

[Write the introduction of a paper on the effects of

climate change on GDP]

Please note that this does not mean that I want to

use Consensus' response via verbatim (copy and
paste) for my work. I want to demonstrate what result
can you get with such simple and straightforward
prompt in Consensus GPT. The following screenshot
details Consensus' response.

279
As can be seen in the following screenshot, which is
a continuation form the previous screenshot,
Consensus suggest research articles that it deemed
relevant with our prompt instruction. You can click on
the author names for these academic papers and as
Consensus has created a link for these academic
articles, you can view these academic articles.

Next, I will demonstrate more ways in using

Consensus. Note that this are suggestions based on
my creativity in completing certain task. You can be
more creative than this and experiment with other
custom GPTs that you deem suitable for your
intended task.

280
I start my query to Consensus with a more detailed
prompt as compared to my previous prompt. The
prompt is:

[Can you find me relevant, current and high impact

academic papers on this topic: The effects of climate
change towards economy. Please suggest papers
from the year 2023 only]

In this prompt, I have specified my search for

literatures from the year 2023 only. This is very
recent at the time of writing this book (end of 2023). I
did this to ensure the currentness of the literatures
and to demonstrate Consensus' capabilities.

281
This prompt yielded five academic papers and you
can click on the links to get access to these papers.
Even if some of the papers are from subscription
based journals, you can still get access to its
information and abstract. You can then retrieve the
literatures using your university's database
subscription. We are using Consensus as initial
search only for the literatures. You can then snowball
to get more literatures, either manually or by using AI
assisted method.

282
I can ask for more literatures as demonstrated in the
next prompt:

[Can you suggest five more literatures]

Consensus then suggest five more academic articles

from the year 2023 as can be seen in the following
and the previous screenshot. I will use academic
article number five as the example for the next
screenshot to better illustrate what can be done with
these information.

283
The following screenshot shows what to expect when
you click on the link provided by Consensus for each
academic article. You will be directed to Consensus
web and the information on the particular academic
article will be displayed there. You can also opt for
the full text via the publisher or access towards the
article using Semantic Scholar as illustrated below.

284
I opt to access the full text via the publisher, in this
case Elsevier from the ScienceDirect database. As
can be seen in the following screenshot, this is not an
open access academic article. As I have
demonstrated in the previous chapters in this eBook,
you can retrieve this academic article using you
university's database subscription. This demonstrates
that Consensus does not only search for open access
literatures, but also have access to search for
subscription based academic articles and literatures.

285
I then can manipulate the information given by
Consensus based on these ten literatures that it has
suggested for my search query. I can arrange the
information in a more organized and presentable
manner as can be seen in the following prompt:

[Can you summarise the findings of these 10 papers

into one literature matrix table? Generate the table in
an editable Ms Word format (layout landscape,
margin narrow).

Using this prompt, I can be creative by asking for a

literature matrix table. I even specified it to be done in
an MsWord file document so that I do not have to
copy and paste the table from ChatGPT to my Ms

286
Word. I even can specify the layout of the literature
matrix table (in this case landscape) and also asks for
the margin of the document to be narrow (in order for
better representation of the table accounting for the
information that will be included in the table).

Consensus than performs the task and provide me a

link to download the literature matrix table.

287
I then can have access to the literature matrix in Ms
Word document file. I can do so much more with this.
I can add columns, request for numbering, edit the
table's design and then further prompt ChatGPT via
Consensus GPT to use the literature matrix that I
have edited as a template for subsequent inclusion of
academic article information.

Note that this literature matrix table serves as a

demonstration purpose only. Never should this be
used as your final production of literature matrix table,
however, you now have a base to work from. This is
just arranging the search done by Consensus in a
more organized manner. If you wish to have a
comprehensive and accurate literature matrix table,
you can upload all of the academic articles' full text
into ChatGPT and prompt it to extract the information
from these academic articles.

To ensure more accuracy of information, you should

carefully read all the academic article and then add
the information that was not captured by ChatGPT
manually into the literature matrix table. This is the
best method to approach literature writing. Only use
AI to assist, not to automate.

288
I can do more based on the information that I already
have. I can ask ChatGPT via Consensus GPT to
produce writing drafts for me as can be seen in the
following screenshot using this prompt:

[Using references from these papers, (please use all

10), write a compelling literature review introduction,
introducing background of the study and highlighting
the centrality (importance) of the topic. Use
professional, academic English equivalent to a top
tier academic journal article]

289
ChatGPT then produces the draft that I can use to
see the flow of argument, suggested structure of
writing and many others. However, I do not feel that it
is a good draft. I can prompt ChatGPT to produce a
better draft. Hence, that is what I did in the following
prompt.

290
I then further prompted ChatGPT to produce a better
draft. An example can be seen using this prompt:

[Please do not use ChatGPT associated vocabulary

such as: realm, multifaceted and others. Assuming
the role of an expert academic article writer, rewrite
the background of study for the literatures. Be
concise and observe coherence throughout the text.
Use effective cohesive devices. Embed the literatures
within the writing using APA style citation (in text
citation), and at the end of the writing, provide the
references for these articles].

291
The following screenshots illustrates the comparison
of prompts results based on text output. You can
compare between the two previous prompts
requesting for the same task and see that the second
prompt yielded better result. The key is to be detail
with your prompts.

Harnessing Custom GPT from GPT Store for

literature review tasks.

292
293
294
295
296
297
298
In this bonus chapter, we have explored the
innovative capabilities of custom GPTs, particularly
Consensus, in aiding the process of literature review
writing in academic research. The demonstrations
provided offer a glimpse into the potential of AI-
assisted methodologies in organizing and
synthesizing information from multiple sources.
These examples, showcasing the utility of GPTs in
searching for relevant articles, creating literature
matrices, and generating initial drafts for the
background of a study, are indicative of the significant
strides made in integrating AI tools into academic
writing practices. However, it is imperative to
understand that these demonstrations are
foundational guides meant to enhance, not replace,
the meticulous and critical processes intrinsic to
academic research.

Academic integrity and rigor form the base of

scholarly writing and research. In the context of
literature reviews, this necessitates a thorough
reading and comprehensive understanding of each
academic article. Researchers must engage deeply
with the content, extracting pivotal information and
critically analysing various perspectives. The
grouping of papers according to themes and the
subsequent synthesis of information is a nuanced
299
process that AI tools can assist with but cannot fully
replicate. The essence of academic writing lies in the
ability to discern, interpret, and present information in
a manner that is both informative and original. While
AI tools like custom GPTs can streamline certain
aspects of this process, they are supplements to, not
substitutes for, the intellectual rigor and analytical
skills required in academic research.

Conclusively, as we embrace the technological

advancements in AI and its applications in academic
writing, it is crucial to maintain a balance. While these
tools offer significant assistance in managing and
organizing large volumes of information, the ethical
implications and quality of academic work must
remain at the forefront. Drafts produced with the aid
of AI should be viewed as initial steps in the writing
process, requiring further refinement, critical thinking,
and personal insight. As researchers and
academicians, our goal should be to use these tools
responsibly, ensuring that our scholarly work upholds
the highest standards of academic excellence and
integrity. Doing this ensures we not only preserve the
sanctity of academic research but also enrich it with
the thoughtful integration of technology.

300
C O N C L U S I O N
The literature review stands as a testament to the
scholarly diligence and intellectual rigor of the
academician. Just as a seasoned chef carefully
selects each ingredient, ensuring that only the finest
make their way into the culinary creation, a
researcher must handpick each study, theory, and
source to weave a comprehensive narrative that
underpins their work. The use of AI in this meticulous
process, as detailed throughout this book, serves as
an able assistant, enhancing the researcher's
capability to identify and collate the necessary
academic 'ingredients'. However, it is imperative to
acknowledge that AI is a facilitator, not a replacement
for the researcher’s critical engagement with the
literature.

The cautionary stance towards the automated

aspects of literature review cannot be overstated. The
richness of a review does not solely rest in the
aggregation of data and sources but in the critical
analysis and synthesis that a human mind brings to
the table. AI tools can sort, organize, and even
highlight patterns in the literature, yet they lack the
capacity for critical thinking and the holistic
understanding that comes from years of scholarly
practice. Literature review writing is not just about

302
reporting what has been found; it is about
questioning, critiquing, and contributing to the
conversation within a field. It is an exercise that
promotes and hones the researcher's critical thinking
skills, ensuring that their work not only informs but
also challenges and advances the scholarly
discourse.

Moreover, the literature review serves a greater

purpose than just a preliminary step in the research
process; it is a demonstration of the researcher's
expertise and their unique perspective on the topic.
The human element of creativity, insight, and the
ability to draw connections between disparate studies
is something that AI, in its current form, cannot
replicate. The review, written by the researcher,
becomes a scholarly voice in the academic dialogue,
positioning their forthcoming work within the broader
context of the field. It is an intellectual craft that
marries the methodical with the imaginative, a skill
set that AI is yet to master.

In conclusion, while AI provides an invaluable

resource in aiding the literature review process, the
soul of the review remains the intellectual imprint of
the researcher. The essence of the literature review

303
is captured not just in the comprehensive coverage of
existing literature but in the insightful interpretation
and thoughtful critique that a researcher brings to
their narrative. As we employ AI to assist us in the
preparatory phases of our academic cooking, let us
not forget that the true art lies in the personal touch,
the seasoning of the dish, which defines the flavour of
our scholarly contributions. Therefore, as we
embrace the future of research with AI as our sous-
chef, we must continue to nurture and value the
irreplaceable qualities of human inquiry, critical
analysis, and intellectual synthesis that are the
hallmarks of an impactful literature review.

The video of the conclusion for the eBook can be

accessed using the link below.

eBook Conclusion

304
 NOTES

As AI is advancing everyday, I feel it is

only logical to provide continuous update
on the recent AI tools or features that can
be beneficial in literature review process.
You can access this update from here.

https://drive.google.com/drive/folders/1EX
nymavNr1O36bOhNyoReOX3_2w4FLo1?
usp=share_link

Thank you
Discover how to craft a compelling literature review with "AI-Assisted
Literature Review." This guide is your key to unraveling the complexities of
literature review in academic research. The eBook offers straightforward
advice complemented by embedded video tutorials to enrich your
understanding, making each step in the process engaging and easy to grasp.

Tap into the capabilities of AI to refine your literature review, with practical
tools that support thorough analysis and efficient organization. The content is
interactive and presented in a clear, user-friendly manner. Ideal for both new
students and experienced scholars, this eBook provides a hands-on approach
to perfecting the literature review, showcasing how technology can be a
powerful ally in the pursuit of academic knowledge.

About the Author

Faharol Zubir is an English language lecturer from
the Department for Languages and General Studies,
Universiti Malaysia Perlis. He is also a research fellow
for the Centre of Excellence for Unmanned Aerial
System UniMAP, certified translator by the Malaysian
Institute of Translation & Books (ITBM) and a
seasoned Malaysian University English Test (MUET)
Examiner. He specializes in thesis content analysis.
With a deep passion for research, he understands the
challenges of thesis writing. He leads specialized
training sessions, aiding both new and experienced
researchers. He had authored the book “ChatGPT for
Thesis Writing”. His recent book, "AI Assisted
Literature Review Writing", showcases his expertise,
blending invaluable insights with AI's potential to
transform academic writing particularly literature
review process and writing. His dedication to the
academic field is clear as he seeks to influence both
local and global audiences.