CHAPTER TWO

Review of Related Literature

2.1 Introduction

This review is on the perception of the implications of Artificial-Intelligence

of use in English Language learning in universities. A thematic approach was taken

for this presentation to provide a comprehensive examination of key areas of the

language learning. The review presents the conceptual framework, theoretical

framework, review of previous studies, and summary.

2.11 Artificial Intelligence: Evolution and Types

Octopeek (2018) notes that in the 19th century, the main applications of

artificial intelligence were in the fields of speculation, imagination, and narrative.

The goal of classical philosophers was to bridge the gap between humans and

machines. But at the time, artificial intelligence was only depicted in fiction, such

as Mary Shelley's "Frankenstein." In 1956, artificial intelligence actually got its

start. He explains further that the beginning of an era controlled by artificial

intelligence was marked by the recognition of attendees at a workshop organised

by Darthmod College as the future leaders of AI. An examination of the last 60

years shows how artificial intelligence has changed since its inception.
 John McCarthy held the first AI conference and first used the phrase

"artificial intelligence" in 1956. He is the father of Artificial Intelligence

(Rajaraman, 2014). Shaky was the first all-purpose robot constructed in 1969.

Rather than just following instructions, it can now carry out tasks with intention. In

1997, the supercomputer "Deep Blue" was developed and managed to defeat the

world chess champion in a match. IBM's development of this large computer

marked a significant achievement. In 2002, the first robotic hoover cleaner was

introduced, marking a significant milestone by achieving commercial success.

From 2005 to 2019, various technological advancements such as a dancing robot,

robotic process automation (RPA), speech recognition, smart houses, and other

breakthroughs have been introduced (Syakis, 2019). In 2020, during the initial

stages of the SARS-CoV-2 (COVID-19) outbreak, Baidu introduced the

LinearFold AI algorithm to support medical and scientific institutions engaged in

vaccine development. The method is 120 times swifter than existing approaches,

accurately forecasting the RNA sequence of the virus in a mere 27 seconds.

There are many different approaches and uses for artificial intelligence,

which fall into the following categories: Training an AI model to find patterns and

insightful information in data is known as machine learning (Hu et al., 2019).

Artificial intelligence employs a variety of techniques, including supervised,

unsupervised, and reinforcement learning. These techniques are flexible and can be
used in a number of domains, including audio and image recognition, natural

language processing, fraud detection, and recommendation systems.

Deep learning is a branch of machine learning that learns and extracts

intricate information from data using multi-layered neural networks. Natural

language processing, image categorization, and speech recognition are a few

examples of applications. In the field of natural language processing (NLP),

computers undergo training to understand and interpret human language

effectively. AI has a variety of applications such as language translation, chatbots,

virtual assistants, and sentiment analysis. Robotics involves leveraging AI

technology to streamline processes that would typically require human

involvement, enabling the operation of robots. Possible applications encompass

home automation, healthcare, logistics, and manufacturing. Computer vision

involves utilizing artificial intelligence to analyze and understand visual data such

as images and videos. Maxime Comptier mentioned that artificial intelligence has

enabled an increase in human resources (Octopeek, 2018).

2.2 Artificial Intelligence App Use in English-Language Learning

Using AI applications for teaching is gaining recognition in experimental

research. According to Fitria (2021), AI education applications are getting more

diverse, making use of both programmed learning and cutting-edge open-source

software. The value of these AI educational applications is in their capacity to meet

learners' needs and abilities, adjust to their chosen learning modes, and track each

student's progress. These apps feature tracks that are tailored to learners' varying

attention spans, encourages them to study, and is inclusive of all skill levels. They

offer feedback that evaluates student performance and points out the advantages

and disadvantages of the scientific information (Devi et al., 2020).

The curriculum's integration of courses makes sure that each topic is logical

and that students are proficient before moving on to more difficult areas. Students

can be taught scientific information by giving them problems to solve that relate to

their particular fields of study. This procedure is monitored by instructors, who

offer direction and criticism. Kamuka (2015) opined that Artificial intelligence

(AI) tutoring systems have the potential to replace teachers in educational settings

by allowing students to adopt their own study techniques and automatically

provide guidance. Cautela et al. (2019) claim that there is a connection between

different AI applications and cognitive skills—like design—that are in high

demand in the twenty-first century.

The goal of teaching non-native English speakers English is to increase their

speaking, listening, reading, and writing skills by expanding their vocabulary and

linguistic understanding (Renji & Thangarajathi, 2024). According to the National

Policy on Education (2014), the goal of teaching English in Nigeria is to provide

them with permanent literacy and the capacity to communicate effectively. This is

why teaching English to non-native speakers is deemed crucial in today's culture.

This also involves employing language to create and understand documents.

Al Mukhallaf (2020) states that as the final product and the foundation of

language development, communication is essential. He explains further that

integrating conventional and digital communication channels into teaching and

learning activities is therefore critical. He suggests that to create language-based

educational games, replicate real-life English conversations and settings, and offer

practical language training, artificial intelligence (AI) applications such as

simulation and communication programs must be included. With the use of audio

and visual assistance, AI-powered communication tools facilitate the creation of

learning environments that enhance language proficiency and pronunciation (Al

Mukhallaf, 2020).

In addition, AI applications allow students to practice their language skills

and receive feedback through comments. Some programs incorporate language

exercises to help learners achieve proficiency by providing instruction in

communication using language abilities (Barnes, 2016). According to Radwan

(2017), artificial intelligence (AI) can be utilized to tackle various obstacles

encountered in teaching and learning English. These obstacles include utilizing

machine translation to enhance students' translation skills, text-to-speech for

visually impaired individuals, online digital language dictionaries to broaden

vocabulary, smart programs to enhance the speaking skills of English language

learners, and the implementation of writing evaluation methods to instruct students

on forming paragraphs and essays (Al Mukhallaf, 2020).

On the other hand, many scholars (such as Huang et al., 2023; Godwin-

Jones, 2022) have pointed out that while there are many benefits of using AI-

powered technology in English language instruction for both teachers and students,

there are also important drawbacks that should not be disregarded. The growing

reliance on technology is one of the main issues as it may weaken human ties and

negatively affect the social and communicative components of language

acquisition. According to Khanzode and Sarode (2020), this problem is thought to

be the primary drawback of AI-assisted technology. Concerns regarding possible

data abuse are raised by the fact that AI systems gather data on how instructors and

students utilise platforms. Therefore, maintaining privacy is crucial to preventing

unauthorised sharing of users' sensitive data. Additionally, since not every student

has the same degree of access, it is imperative to guarantee equitable access to new

technology and the internet. Thus, if AI language learning tools are used in

language training, it is crucial to make sure that teachers and students can utilise

them with ease. While some AI applications allow users to converse in real time
with English speakers or language teachers, the majority of these technologies are

largely focused on self-directed learning with no human involvement.

In addition to the challenges associated with the use of AI applications, there

are ethical considerations to take into account when incorporating AI tools in

English language education. One such concern is the difficulty of understanding

and reproducing language nuances like idiomatic expressions and regional accents

(Font de la Vall & Araya, 2023), which can lead to mistakes or awkward language

usage. AI tools may also not be as effective at detecting errors and providing

explanations as human teachers are, which could result in students repeating

mistakes or developing bad language learning habits.

Edmett et al., (2023) also noted another challenge in the intergration of AI

apps in English language learning. The scholars stated that the standardisation of

philosophies and languages was one of the biggest obstacles as these were

consistently changing. Rowe (2022) found that while Google Translate's

programming tended to ignore the subtleties in various language groups, it seemed

to communicate signals about suitable and standardised language use for pupils in

a second-grade American school. A student gave the example of discovering that

the only language choice available for translating Tagalog to English on Google

Translate was Filipino, which has been the officially recognised language of the
Philippines since 1987. Tagalog was not included as a language option. According

to Rowe (2022), this circumstance made the student wonder what language is, who

has the right to define it, and whose language is "correct." This suggests that the

adoption of standardised language may be encouraged by Google's preference for

specific historical and political language limits.

According to Dugosija (2024), the use of AI in English Language Teaching

(ELT) has various advantages but it also presents significant difficulties. He noted

that AI technology advancements offer customized learning experiences,

immediate feedback, and improved language acquisition opportunities for students.

These innovations assist teachers in addressing different learning preferences and

individual student requirements, resulting in a more effective and engaging

learning setting. AI tools support educators in tasks such as creating learning

resources, assessing evaluations, providing feedback, and enhancing language

skills through interactive applications. Despite the advantages, there are obstacles

to integrating AI in ELT. Ethical issues, like data privacy, access to AI tools, and

incorporating AI tools in writing tasks, are significant hurdles. There is also a risk

of relying too heavily on technology, which could reduce human interaction and

the cultural, emotional, and social aspects of language learning. To fully utilize AI

in the teaching of English language and mitigate potential risks, it is essential to

adopt a balanced approach. Language educators should view AI as a complement

to traditional teaching methods rather than a replacement, integrating the benefits

of AI with their expertise.

2.3 AI Tools and their Adaptability to English Language Teaching and

Learning

According to Fitria (2021), AI improves the effectiveness of the English

language learning process and helps spread a variety of knowledge. These pupils'

understanding of the teacher's explanations is facilitated by the availability of many

types of educational technology. Even if they don't have direct communication

with their lecturers, students may still learn with ease. There are several artificial

intelligence (AI)-based language learning applications available for use by both

English teachers and students/learners. Examples of AI technology that can be

used to improve English learning include the following:

i. Google Translate:

Google Translate's role as an online translator is crucial when it comes to

teaching and thinking about English. Understudies can use it to comprehend words,

phrases, sentences, chapters, and other elements that are frequently used in

Indonesian English interpretation exercises. Learners can use Google Translate to

examine the spelling of words that appear due to errors. It is helpful when it comes

to verifying spelling in English. Google Interpret is one tool for learning how to
pronounce words in other accents. Understudies who want to think about English

for complementary purposes, particularly in terms of word articulation, might use

Google Decipher (Covilli, 2016).

ii. English ABLE

The learning environment known as English ABLE employs evaluations to

concentrate on English grammar. Zapata-Rivera and his group created the method.

In 2007. Assessment data from several sources, including formative and

summative tests, informs instructional guidelines in assessment-based learning

environments (ABLE). In order to help English language learners (ELLs) master

English grammar during class and study sessions, English ABLE offers an

educational environment that places a high priority on evaluations. Using a TOEFL

CBT question bank, English ABLE has created new collections with challenging

tasks catered to various English language learners' skill levels. Additionally, a

flexible and encouraging learning environment provides tools to help students

succeed in a variety of English grammar domains (Fitria, 2021).

iii. Orai

Orai is the ideal choice for improving your public speaking abilities. It is

easy to use. Improving your public speaking abilities is a quick and self-directed

process. Orai is excellent at determining the number of words and filler words used
in English instruction and learning. Teachers can easily incorporate Orai into their

classroom activities by incorporating English-speaking topics like character

descriptions. After giving students a convenient fifteen minutes to use Orai, a

student can be randomly selected to come forward and practice speaking right

away. Orai offers a variety of tools, including Lessons, Practice, Progress, and

Recordings, to help students improve their speaking abilities.

2.4 Teachers’ Perceptions of Implications of Artificial-Intelligence App Use

in English-Language Learning

Language learners have been using Machine Translation extensively for over

a decade (Clifford et al., 2013). According to Ata and Debreli (2021), 94% of

students studying English as a second language use MT. Among the most well-

known uses of MT tools are the lexical meaning of a single word (Briggs, 2018;

O'Neill, 2019), constructed interpretations (Ata & Debreli, 2021), elocution

assistance (Ata & Debreli, 2021), task completion (Briggs, 2018), and

comprehension (O'Neill, 2019). Since Anderson first suggested using MT as a tool

for L2 learning in 1995, CALL research has taken into account the idea that L2

teaches almost MT. Despite the widespread use of MT by students, the majority of

dialect teachers detest it (Clifford et al., 2013).

 The two main issues that instructors face when using machine interpretation

(MT) are detecting scholastic dishonesty and using MT to evaluate dialect

proficiency (Ducar & Schocket, 2018). According to Henshaw (2020), educators

have employed a variety of approaches to deal with MT over time, ranging from

including interpretation exercises into their classes to outright banning it. Some

researchers have examined teachers' perceptions of chatbots. In agreement with

Chocarro et al. (2023), instructors are more likely to use them if they believe they

are acceptable and simple to use. Pokrivcakova (2022) surveyed aspiring English

instructors within the Czech Republic and Slovakia with respect to their

information of and suppositions on the utilize of chatbots for educating the dialect.

Instructors gave these devices tall marks for qualities like "fun," "exact," and

"engaging," and more than one-third of respondents expressed they would

consolidate chatbots into their lessons, whereas about half said they would not.

Comparable to Pokrivakova, Belda-Medina & Calvo-Ferrer (2022) examined how

future dialect teachers would see the utilize of conversational AI in dialect

learning. They found that whereas members communicated intrigued in learning

more around chatbots, they showed up to esteem human-to-human communication

over human-to-chatbot communication.

According to Kasneci et al. (2023), LLMs could negatively affect students'

motivation, critical thinking, and problem-solving skills. They recommend tactics

like using LLMs to teach students about their obstacles and to essentially analyze

content related to AI; using AI to investigate some theories instead of advertising

plans; and, more generally, integrating AI in a way that improves the educational

plan. They also advocate for educators to use AI as a supplement, arguing that they

can become overly dependent on technology. Teachers have found it more difficult

to distinguish between works created by humans and machines (Cotton et al.,

2023; Elkins & Chun, 2020).

Despite the fact that these strains are real, they may also indicate that

policies and evaluation methods need to change in order to keep up with

innovations. The use of AI in education may be a complex topic, especially when it

comes to academic judgment. Perkins (2023) discusses how LLMs such as

ChatGPT might be useful tools for assisting understudies with their research.

However, because the LLM outcome is so coherent, other algorithms, like the

academic copyright infringement discovery software, frequently fail to identify it.

Rules pertaining to understudy use must be properly drafted by colleges. This

situation suggests that the tensions around AI may be partially caused by

traditional standards and techniques. Additionally, certain plagiarism-detection

technologies "exhibited inconsistencies, produced false positives and uncertain

classifications" (Elkhatat et al., 2023).

 Research by Hwang and Cheng (2023) on how teacher educators perceive

the possible effects of GenAI on curriculum, teaching, and assessment confined

that they understand that for language instructors to succeed in the curriculum,

GenAI will demand adjustments to the competencies and abilities they now

possess. As a result, ILTE programs will have to update and modify their

instructional learning objectives. Teacher educators brought up important issues

about language learning requirements, the function of language teachers, and

language acquisition skills. It is generally accepted that maintaining traditional

teaching techniques will result in lower pupil motivation and a smaller role for

instructors. The potential for GenAI systems to handle a variety of language-

related activities, such simultaneous translation, might make language learners less

motivated to learn new languages. Moreover, the notion that these instruments can

serve as mentors for language acquisition (Hwang & Chen, 2023) may lessen the

necessity for conventional human-centered instruction. In early 2023, however,

just two months after ChatGPT was made widely available, a study by Ali et al.

(2023) discovered that students and language teachers thought that ChatGPT-based

instruction could have a positive effect on language learners' motivation to learn

independently, intrinsically, and extrinsically. Teachers understood how crucial it

was to show students how to utilise GenAI technologies and to have meaningful
conversations with them about how these tools will affect their education and

future professions.

The several viewpoints demonstrate how AI is being implemented in

education in a way that is always changing. Although academics acknowledge the

benefits of AI, they are also concerned about its drawbacks, which highlights the

need for a cautious and moral approach. According to Abdelaal and Sawi (2024),

assessing AI's possible advantages requires a thorough grasp of its effects.

Practices that are equitable, moral, and student-centered are necessary for the

successful integration of AI in education.

2. Students’ Perceptions of the Implications of Artificial-Intelligence App Use

in English-Language Learning and Practice

AI chatbots and intelligent voice assistants, like Google Translator or Alexa,

have been shown to help students become more proficient in conversations and

less nervous while speaking a distant dialect (Tai & Chen, 2023). According to

Bao (2019) and Tai & Chen (2023), AI chatbots that aim to develop conversational

skills can reduce learning barriers and conversational anxiety. Dizon (2020)

discovered that while understudies made significant improvements in their talking

ability, they did not advance their tuning in abilities in a similar manner. El Shazly

(2021) discovered that while AI chatbots might be useful in improving phonetic

pick-up, they did not lessen learners' speech-related tensions. In a similar vein,

Çakmak (2022) discovered that students' anxiety levels rose when interacting with

chatbots since they had a hard time being understood. According to a study by

Hew et al. (2022), chatbots can help EFL students create goals. They discovered

that students thought the chatbots were practical and simple to use (Zimotti et al.,

2024).

Undergraduates also perceive AI's use in education favorably since it offers

so many benefits (Chan & Hu, 2023). According to Branson et al. (2019), they

include improving the quality of education, expanding the quantity of learning

opportunities, and perhaps affecting the understudy experience. Through the use of

fake insights, each student could have a customized learning partner who would

not only act as a learning partner, coach, and teacher but also continuously record

the student's progress and interface (Holmes et al., 2019), improve their

communication skills, and help them develop their tuning in abilities (Mohammed

Mahmoud Ghoneim). Learning execution can also advance when students

understand how to use innovation (Moghavvemi et al., 2023). In their study,

Azzahra et al. (2024) observed that students saw AI-based language learning

applications as instruments to improve language learning. According to the

participant, programs such as Duolingo and ELSA Speak have produced

immersive experiences by integrating a variety of interactive elements that can

mimic real-life communication. These results also demonstrate that English

education students have a positive opinion of AI-based language learning

applications because they can successfully enhance language acquisition,

particularly speaking and listening abilities. These findings are consistent with

those of Cherner et al. (2023).

In Ningsih et al.‘s (2024) study, they note that the majority of Indonesian

college students now exhibit knowledge of artificial intelligence and its many uses.

According to the findings of the variable analysis, the participants think AI

significantly improves their motivation to study and helps them reach their learning

goals. Since the AI-driven apps are practical and satisfy their demands, they have a

favorable attitude toward the inclusion of AI in English learning courses. The

encouraging results are consistent with other research showing that AI enhances

students' vocabulary and pronunciation while also preparing them for interactions

in English as a Foreign Language (EFL) (Summakul et al., 2022). The majority of

learners agree that it makes a difference to improve dialect competence and reduce

anxiety because the devices vary and some of them are already identifiable with

AI. With the aid of Duolingo and BBC Learning English, they significantly

improve their elocution and speaking abilities. The majority of understudies use

Chat GPT, Grammarly, and Quill Bot for data extraction and academic writing. In
addition to the citation management system, they use Mendeley and Zotero to

organize their work (Ningsih et al., 2024).

Using AI tools can improve university students' English writing abilities

considerably without impeding their language development (Gayed et al., 2022;

Klekovkina & Denié-Higney, 2022). According to O'Neill and Russell (2019),

university students responded favourably to Grammarly's grammatical

recommendations and said the program was helpful and user-friendly.

Qassemzadeh and Solimani (2016) found that students retained passive voice

grammar norms better when they got feedback on sentence creation from AI

systems rather than professors. Reis and Huijser's 2016 study was more evidence

that Grammarly offered thorough feedback and useful tools for writing in English.

This implies that the ease of use as well as the provision of feedback integrated in

AI apps aids the language learning process.

About half of Korean ELLs thought online machine translation (MT) was a

useful tool for learning the language, according to Briggs (2018). The benefits of

automated AI feedback systems for Korean English language learners in academic

settings have been the subject of earlier research, which this study supports (Lee,

2020; Lee & Briggs, 2021). Jeong's (2021) research revealed that English language
learners (ELLs) had an enhancement in their writing proficiency while utilising

automatic translation tools like Google Translate (GT) and NP. Furthermore, a

great deal of English language learners (ELLs) in university settings used

automatic translation systems to support their language acquisition. This led to

advancements in vocabulary, grammar, and expression, as well as the editing of

English papers. Chen et al. (2023) compared the attitudes of English Language

Learners (ELLs) towards AI-based technologies with the perceptions of Google

Assistant (GA) capabilities at varying competence levels. Using a range of voice

commands with GA for an hour, 29 college students studying English as a second

language participated in the study. The participants said GA was a helpful tool for

language acquisition, particularly for speaking and listening skills, and they

enjoyed interacting with it. The participants also saw that GA pronounced words

clearly and organically. With the use of GA, more proficient students were able to

improve mutual comprehension, while learners at lower proficiency levels faced

more difficulties as a result of mispronounced words. These findings applies in the

Nigerian context English language, according to Nnyigide (2020), is also a second

language in Nigeria.

The attitudes, experiences, and perspectives of EFL students about the use of

AI technologies or apps in English language instruction have been the subject of

several research carried out in a variety of circumstances, and these studies have
yielded insightful information (Darwin et al., 2024). Although most people

acknowledged the advantages of AI technology, opinions on their efficacy and

adoption varied. According to Ghafouri's (2024) research, ChatGPT has the

potential to improve L2 persistence and facilitate communication between Iranian

EFL teachers and students. In her investigation into the opinions and experiences

of EFL students using Grammarly, an AI-powered tool, Alotaibi (2023) discovered

that the students valued Grammarly's capacity to spot grammatical errors and offer

advice on how to improve the calibre of their written work. Notwithstanding these

benefits, the study also brought to light concerns about potential overuse of the tool

and meaning changes, which may impede students' development as writers (Sol et

al., 2024).

How Perception of AI App Use influence English Language Learning

Technology is advancing so quickly that an AI will fit into the conventional

classroom, creating a different environment that may become more productive than

the conventional classroom setting. The study's findings demonstrated that

instructors' decisions to stick with using AI technology in their lessons are

influenced by their perspectives. AI technology will inevitably be integrated and

used, and prior research indicates that it will have a big influence on the education
industry down the road. The methods used to teach English and English as a

second language are impacted by the fact that English is a language that is spoken

around the world. Teachers play a critical role in helping students learn in the

classroom, thus it's necessary to take into account their goals about the use of

artificial intelligence and their opinions on the technology. In the digital age,

Artificial Intelligence in Education (AIEd) has demonstrated a strong capacity to

modify the teaching and learning procedures to meet the needs of today's pupils.

This is in line with the goals of the UNESCO technology Innovation in Education

(2021–2025) plan, which promotes inclusion and justice while ensuring that

everyone has an equal chance to profit from technology breakthroughs, particularly

in innovation and knowledge (Zulkarnain & Yunus, 2023).

When it comes to students' confidence in utilizing digital media for learning,

studies have revealed that students with high internet self-efficacy often do better

on online learning activities and have better search methods than students with low

internet self-efficacy. Students with poor self-efficacy may be reluctant to explore

novel approaches to information retrieval and online problem-solving, in contrast

to students who are confident in their internet-related skills.

According to studies, students' degree of involvement in technology-

integrated learning settings is largely determined by their belief in their

technological prowess, and their confidence in using computers can affect both
their engagement and performance in the classroom. In a study conducted by Chai

and colleagues in 2021, it was found that self-efficacy was the primary factor that

accurately predicted students' intention to engage in certain behaviors. Moreover,

students' attitudes towards the use of artificial intelligence (AI) for positive social

impact, such as solving problems and enhancing people's well-being, and their

willingness to learn AI were notably influenced by their perceptions of the benefits

of learning AI for social good (Baez et al., 2023).