Running Head: Proposal

A review of the Artificial intelligence on decision making

processes in stock valuation
[Name of Student]

[Name of Institution]
 2

Table of Contents
Introduction......................................................................................................................................3
Background and Rationale...........................................................................................................3
Project Aim and Objectives.........................................................................................................4
Significance of Study...................................................................................................................4
Research Question:......................................................................................................................5
Literature Review............................................................................................................................6
Theoretical Framework:...............................................................................................................6
AI in Decision-Making for Stock Valuation:..............................................................................6
The Role of AI Techniques and Methodologies:.........................................................................7
AI Effectiveness and Limitations:...............................................................................................8
Research Methodology....................................................................................................................9
Research Philosophy:...................................................................................................................9
Research Design........................................................................................................................10
Data Collection..........................................................................................................................10
Data Analysis.............................................................................................................................10
Project Schedule............................................................................................................................11
References......................................................................................................................................12
 3

Introduction
Background and Rationale
AI has significantly impacted the financial market regarding decision-making
mechanisms, as well as in many other spheres. As for stock valuation, AI has extended from
simple machine learning approaches to using neural networks and other deep learning techniques
that have changed traditional modelling paradigms (Pickover and Menasche, 2019). They allow
financial institutions to perform large numbers of calculations on large volumes of data quickly
and efficiently, thus reducing the need for a workforce and improving predictive models. Hence,
traders and investors can easily exploit the real-time data to make decisions rapidly and adopt
appropriate strategies in fluctuating market rates.

The reason for selecting decision-making related to stock valuation as an area for
applying AI is due to the high efficiency of AI in analysing and interpreting large volumes of
financial data. In contrast to such quantitative methods, which often involve extrapolating past
patterns and statistical outputs, AI-based computations need to be more accurate in pinpointing
new patterns and relationships within datasets that human analysts may not easily discernible
(Jacob and Nerlinger, 2021). As such, this capability is invaluable in complex and informational
environments where quick and accurate decision-making is required to exploit market
opportunities and avoid or mitigate possible risks.

Although the application of AI to assist in the financial decision-making process holds

much value, the current literature reveals several difficulties in integrating such systems. Some
of the issues are as follows: One critical challenge is the reliability of the AI models in different
markets and managing unexpected events (Siau and Wang, 2020). A glance at AI models reveals
that they are highly accurate predictors, but improving the AI models' robustness is essential.
Ethical issues also come to the foreground regarding determining information based on artificial
intelligence. Some of the central questions that concern us include data protection, algorithmic
fairness, and the societal impact of AI decision-making. The continued integration of AI
technologies in the financial markets makes it compulsory to set standards and policies that can
act as a guide in the future.
 4

However, the regulatory environment is another challenge that must be considered while
attempting the entry strategy (Liu et al., 2024). The current and emerging legal and regulatory
approaches to governing AI applications in finance often need to be updated and able to keep up
with technological developments, yet they present new sources of risk and regulatory confusion
to financial institutions. The difficult balance between encouraging innovation and preventing
and punishing bad practices is the key to correctly introducing artificial intelligence into
economic processes. Combating these gaps and challenges is imperative for the progression of
AI-assisted decision-making in stock evaluation. Improving model robustness, mitigating the
risks associated with ethical concerns, and devising strategies for engaging with the regulation
properly will help financial institutions leverage AI's full benefits in economic decision-making.

Project Aim and Objectives

This research aims to critically consider the role of artificial intelligence in decision-
making in the context of stock valuation in financial markets. More precisely, it is intended to
reveal the factors that facilitate or hinder AI in helping to improve decision-making and to
provide insights on how AI might be better utilized in shaping its financial practices. The study
accomplishes the following specific objectives:

 To Conduct a literature review to identify relevant prior research on the use of AI in

decision-making for stock valuation.
 To evaluate the efficacy and drawbacks of AI algorithms in analysing stock prices and
coming up with the best investment strategies.
 To assess ethical questions and legal issues pertinent to using artificial intelligence in
decision-making in financial markets.
 To suggest measures to improve the overall stability, efficiency, and fairness of using AI
to set the stock value.

Significance of Study
The implications of this study are pertinent to academic, practical, and policy use
domains in elucidating the effects of AI in the decision-making landscape of financial markets.

From an academic perspective, this research fits into the emerging literature on AI and
finance by offering overviews of how emerging societies are revolutionizing conventional
 5

approaches to stock valuation (Enjolras et al., 2023). In this way, the study develops theoretical
concepts regarding the AI's capabilities to work with large flows of information and find
complex patterns, as well as expands the knowledge of sophisticated methods used in financial
decision-making processes. Moreover, it serves as a baseline for subsequent research studies
seeking to establish the role of AI within the dynamics of financial markets in the future.

Thus, from the pragmatic perspective, analysing how AI impacts decision-making

regarding stocks is crucial for financial institutions, investment firms, and market players.
Leveraging the capabilities of AI has been shown to enhance accuracy, efficiency, and timely
analyses of stock prices in the modern investment decision-making process (Enjolras et al.,
2023). The knowledge obtained from this analysis can assist practitioners in managing markets
and enhancing the application of AI to address different environments, strengthening the
portfolio methodology and increasing financial results.

On the policy level, the research has captured some of the crucial areas in using ethics in
artificial intelligence in the financial sector, the aspect of data privacy, and the role played by
regulations. Policymakers could benefit from this study and develop proper benchmarks to
validate that the AI processes involved in financial decision-making are both fair and
accountable (Chen et al., 2023). In this respect, the choices undertaken by policymakers whereby
utilisation of further AI technologies in several spheres would be facilitated are crucial while
stressing market value and investor safeguarding.

For financial markets to be devised in the contemporary world, it is crucial to understand

the impact of artificial intelligence, chiefly because technologies and analytics in such financial
markets are imperative. AI is no longer only about cost savings and optimization but also about
changing markets and the way risks and, consequently, investments are managed (Sund et al.,
2016). In this study, all the potential impacts of AI were explored to stress the need to shift to
technologies and manage the opportunities of AI to foster sustainable development and maintain
competitiveness in the global financial markets.

Research Question:
What are the implications of using AI in decision-making regarding stock valuation in
financial markets?
 6

Literature Review
Theoretical Framework:
The theoretical foundations of AI used in financial decision-making involve various
domains essential for comprehending their use in stock analysis. Supervised learning and
reinforcement learning are some of the primary techniques relevant for creating and
implementing AI models capable of processing vast amounts of data (Jacob and Nerlinger,
2021). Supervised learning includes using training data and making the models learn to identify
patterns and make predictions, while in reinforcement learning, the algorithms learn from
feedback and errors received from an environment. Neural networks because they are part of the
Machine learning ecosystem and because of their structure since they are designed in layers of
artificial neurons to make predictions from sophisticated financial data. The major advancement
in intrusive AI techniques has been through neural networks, especially deep learning neural
networks, because they better determine temporal dependency and the non-linear characteristics
of the financial markets.

Extending behavioural finance theories, such as the fear or greed influencing investor and
market decisions, does more than take AI techniques to the next level; they are built on this
foundation. It also complicates the concept of the AI models, bringing in irrationality and other
tricky factors, such as emotional bias in market decision-making. The heuristics and bias by
Jacob and Nerlinger (2021) outlined how irrationality deviates from rationality and the way
Barberis and Thaler described how these biases would impact markets. The theoretical paradigm
draws ideas from areas of machine learning, artificial neural networks, and the behavioural
finance discipline and can lay a strong theoretical platform for applying AI solutions in the
financial market context. Besides, it optimizes time factors of quality and speed of decisions and
deals with major challenges such as interpretability, bias, and compliance (Noyes et al., 2017). In
this context, it is crucial to elaborate on the ideas for future studies on AI for stock valuation,
where the emphasis should be on enhancing these theoretical frameworks to advance the AI
models and their responsible application in a diverse financial setting.

AI in Decision-Making for Stock Valuation:

Using this AI in the financial markets is a significant next step in advancing the
computation of stock values. Initially, academicians attempted to solve this problem through
 7

static approaches of classical machine learning, specifically Support Vector Machines and
Random Forests, with the input of past stock data (Leong et al., 2023). These algorithms used
statistical patterns and the history of stock prices to predict the future movements of stock prices,
forming the base for applying artificial intelligence in the financial industry.

In recent years, significant progress in deep learning methods has changed how AI
technologies approach predicting stock prices. Convolutional Neural Networks (CNN) and Long
Short-Term Memory (LSTM) are two examples of latest-generation network models that have
revolutionized financial data processing. CNNs, developed for image detection, are also used to
extract important features from some of the economic time series data, thus capturing complex
structures that other models might not detect (Talamo et al., 2021). At the same time, LSTM
networks are specifically useful in time series analysis to capture time-dependent correlations of
stock prices over long periods.

Such advancements in deep learning have helped AI models improve the performance of
large sets of data with increased accuracy and speed. Using big data, an AI algorithm can
recognize possible temporal dependencies and nonlinear properties characteristic of the financial
markets. Indeed, this capability increases the accuracy of predicting stock prices and refines the
decision-making process among traders and investors (Jacob and Nerlinger, 2021). However, this
deployment of AI in decision-making concerning stock valuation is wider than just prediction
factors. It enables the real-time evaluation of market changes, risk levels, and portfolio decisions,
thus enhancing financial professionals' decision-making abilities by outlining clear strategies
based on real-time market data. Future studies on AI will likely strategize on improving these
deep learning strategies and counter issues concerning model explainability, reduction in bias,
and flexibility for volatility in the market.

The Role of AI Techniques and Methodologies:

AI tools are essential for enhancing decision-making mechanisms in different fields,
especially for processing structured and unstructured data. An element of this capability is the
network of neurons designated as architectures, belonging to a broad category of the same name,
one of which includes deep learning frameworks (Chen et al., 2023). Two types of such
architectures include perceptron, which is made of several layers of neurons connected so that
the neural network can learn the new data without human interference. Considering the choice of
 8

the model for stock valuation, it is interesting that CNN and LSTM models provide the most
efficient solution. CNNs are useful in extracting spatial features from the data intended for the
object recognition process but applied to the temporal analysis of the financial time series data.
On the other hand, LSTM networks effectively capture dependent relationships in sequences and
are suitable for characterizing the stochastic features in stocks at various time intervals.

Another factor is reinforcement learning frameworks, which are also applied to strategies
for financial institutions. Unlike supervised learning, this type of artificial intelligence enables an
algorithm to learn the best decision-making policy by testing actions within the context of the
system environment, which in the present study refers to the financial markets by Danielson and
Ekenberg (2023). Thus, the reinforcement learning algorithm gets feedback on the outcome of
learned actions while altering the working model to optimize the sum of rewards or reduce the
probability of penalties over time. This flexible update characteristic makes reinforcement
learning most applicable for stochastic and volatile situations, which are common features of
financial markets. It is important to note that these AI techniques are not limited to prediction
methods only. They help make immediate assessments of market situations, potential trades, the
best methods to employ for managing risks, and the best ways of managing portfolios (Pickover
and Menasche, 2019). Furthermore, AI accelerates the efficiency and accuracy of workflow to
process a massive amount of data and contributes to the effectiveness of decision-making
patterns of financial professionals rather than guesswork or past tendencies.

AI Effectiveness and Limitations:

AI models have exhibited vast amounts of efficacy in stock valuation, particularly
because of their capability to process big data and identify relationships that may not be visible
to linear models regularly used by humans (Jacob and Nerlinger, 2021). Some of the advantages
of artificial intelligence include Cre server learning algorithms, deep learning models, and
enhanced predictive accuracy on structured and unstructured data for decision-making. It is
crucial to consider several limitations in adopting AI-driven approaches for promoting health
literacy. Decision explainability is particularly important where AI-made choices and actions
require justification. Neural networks and other complex artificial intelligence models are even
described as 'black boxes,' implying that their inner structure is not easily comprehensible to
 9

users (Siau and Wang, 2020). This has made clarity of how the decisions are arrived at difficult,
affecting stakeholders' trust and acceptance.

Another concern that concerns people we are dealing with includes the pre-programmed
bias found in Artificial Intelligence. Systemic biases appear from the training data for the models
based on past discrimination or errors repeated into the algorithmic prognosis (Pickover and
Menasche, 2019). Biases in this regard can result in decisions that have undesired consequences
or perpetuate other social injustices, particularly within the most vulnerable spheres, such as
financial decisions. AI-driven analysis of the financial markets gets theoretical angles from
behavioural finance that factor in human components in the process. Behavioural finance focuses
on human behaviour and psychological factors that affect financial markets, such as feelings,
prejudices, and illogical behaviour in the market. Combining these insights from social
psychology with AI predictions can reduce the likelihood of adverse effects related to market
herding, overconfidence, or gambles that current AI models may not always capture in sufficient
detail.

Research Methodology
Research Philosophy:
Positivism is the research philosophy adopted by this study since it insists on the
possibility of obtaining practical knowledge through observing natural phenomena scientifically
(Talamo et al., 2021). Regarding this research on AI in financial decision-making, positivism
corresponds to the deductive method of investigation, where hypotheses are developed from
theories and tested through empirical research.

This study shall thus take a positivist research orientation to do an empirical investigation
of the theories holding as to how AI can be used in stock valuation. It aims to offer accurate and
impartial information concerning how artificial intelligent technologies shape the decision-
making process within the global markets (Noyes et al., 2017). The method employed by the
study is intended to provide empirical evidence and systematic observations where it could
provide a perspective into how AI technologies play out in the practices of finance, hence
advancing the knowledge base of practice, policy, and research for the experts, makers, and
scholars in the domains of finance and artificial intelligence.
 10

Research Design
The research strategy employed in this study is the deductive type of research, which
tests confirmed hypotheses arrived at from the AI literature on the effects of AI on the decision-
making process in stock valuation. As a cause-and-effect-based style of reasoning, deductive
reasoning enables the formulation of hypotheses and theories about certain phenomena and
making related predictions, in this case signifying the broad aim of studying the impacts of AI
technologies on stock valuation processes in various financial platforms (Chen et al., 2023).

This method aims to identify key aspects of the efficacy and some issues that may arise
when AI is applied in the economic decision-making process concerning existing theories and
past research findings. Also, an abductive analysis will be used of the collected data during the
research process (Talamo et al., 2021). However, in the context of this study, abductive
reasoning will help to consider the emergent themes and patterns in the literature on AI in stock
valuation and enhance the findings based on this deducing approach with richer context and
intermediary ideas.

Data Collection
The main research technique adopted in this current research study shall be secondary
research, which entails collecting data from existing literature. To begin, twenty-five articles will
be accessed through an extensive search using resources such as the Web of Science and Scopus.
The article selection criteria are the type of journal and years of publication; the articles should
be published in academic journals, the most recent ten years, and should discuss AI and stock
value (Talamo et al., 2021). Abstracts will be omitted in cases where the articles fail to meet any
of these conditions or if they are mere theoretical reflections with little or no substance or
empirical evidence. Selecting the articles will involve sifting through the abstracts and the full
texts to obtain only the most relevant and reliable research-based articles (Pickover and
Menasche, 2019). The final sample of articles to be reviewed will consist of 5 articles pertinent
to the research topic, methodologically sound, and informative about the application of AI in
stock valuation decision-making.

Data Analysis
The thematic analysis shall be used as the research approach for analysing the selected
articles. Thematic analysis is an open and processual oriented approach efficiently used for
 11

coding and finding patterns or themes in the obtained qualitative data. In this study, the mode of
subjects, that is, the content of articles, will be coded to locate themes within the flow of articles
and relate to effectiveness, shortcomings, ethics, and regulation of applying AI in stock valuation
(Siau and Wang, 2020).

The process will commence with the operationalization of the study by reading through
the identified articles and getting acquainted with them. Then, the articles will be coded to
develop the first themes (Danielson and Ekenberg, 2023). These will then be gone through to get
a second opinion and harmonize them into a theoretical framework that neatly articulates the
findings in the literature. Using thematic synthesis, this systematic review wants to give an
overview of current knowledge and prospects of AI applications in the decision-making
processes of financial markets.

Project Schedule

Figure 1: Gantt Chart

 12

References
Chen, L. et al. (2023) ‘The impact of governmental COVID‐19 measures on manufacturers’
stock market valuations: The role of labor intensity and operational slack’, Journal of
operations management, 69(3), pp. 404–425. Available at:
https://doi.org/10.1002/joom.1207.

Danielson, M. and Ekenberg, L. (2023) Real-Life Decision-Making. 1st ed. Milton: Taylor &
Francis Group.

Enjolras, M., Galvez, D. and Camargo, M. (eds) (2023) Decision-Making Tools to Support
Innovation : Guidelines and Case Studies. First edition. London, England: ISTE Ltd and
John Wiley & Sons, Inc.

Jacob, A. and Nerlinger, M. (2021) ‘Investors’ Delight? Climate Risk in Stock Valuation during
COVID-19 and Beyond’, Sustainability, 13(21), pp. 12182-. Available at:
https://doi.org/10.3390/su132112182.

Leong, K.-Y. et al. (2023) ‘Bank stock valuation theories: do they explain prices based on
theories?’, International journal of managerial finance, 19(2), pp. 331–350. Available at:
https://doi.org/10.1108/IJMF-06-2021-0278.

Liu, G.L., Darvin, R. and Ma, C. (2024) ‘Exploring AI-mediated informal digital learning of
English (AI-IDLE): a mixed-method investigation of Chinese EFL learners’ AI adoption
and experiences’, Computer assisted language learning, pp. 1–29. Available at:
https://doi.org/10.1080/09588221.2024.2310288.

Lussange, J. et al. (2019) ‘Mesoscale impact of trader psychology on stock markets: a multi-
agent AI approach’, arXiv.org [Preprint]. Available at:
https://doi.org/10.48550/arxiv.1910.10099.

Noyes, J.M. and Cook, M. (Malcolm J. (eds) (2017) Decision making in complex environments.
1st. Boca Raton: CRC Press.

Pickover, C. and Menasche, S. (2019) ‘Artificial Intelligence’. Tantor Media, Inc.

 13

Siau, K. and Wang, W. (2020) ‘Artificial Intelligence (AI) Ethics: Ethics of AI and Ethical AI’,
Journal of database management, 31(2), pp. 74–87. Available at:
https://doi.org/10.4018/JDM.2020040105.

Sund, K.J., Galavan, R.J. and Huff, A.S. (eds) (2016) Uncertainty and strategic decision making.
1st ed. London, [England: Emerald Group Publishing Limited.

Talamo, A., Marocco, S. and Tricol, C. (2021) ‘“The Flow in the Funnel”: Modeling
Organizational and Individual Decision-Making for Designing Financial AI-Based
Systems’, Frontiers in psychology, 12. Available at:
https://doi.org/10.3389/fpsyg.2021.697101.