IEEE EDUCATION SOCIETY SECTION

Received 7 February 2025, accepted 20 February 2025, date of publication 27 February 2025, date of current version 6 March 2025.
Digital Object Identifier 10.1109/ACCESS.2025.3546319

A Comprehensive Systematic Review of AI-Driven

Approaches to Self-Directed Learning
MUHAMMAD YOUNAS 1, DINA ABDEL SALAM EL-DAKHS1 , AND YICUN JIANG2
1 Language and Communication Research Laboratory, College of Humanities and Sciences, Prince Sultan University, Riyadh 11586, Saudi Arabia
2 School of Foreign Languages, Shenzhen Technology University, Shenzhen 518118, China

Corresponding author: Yicun Jiang ([email protected])

This work was supported APC part by Prince Sultan University under the Language and Communication Research Laboratory under Grant
RL-CH-2019/9/1; and in part by the Youth Fund Project for Humanities and Social Sciences Research, Ministry of Education, China, under
Project 21YJC740022.

ABSTRACT Artificial intelligence (AI)-enhanced technologies are increasingly being researched and
applied in learning and instruction. Previous studies have demonstrated the growing influence of AI tools
in improving learning outcomes and transforming the educational landscape. However, limited research has
explored learners’ engagement with AI resources to support self-directed learning, and no comprehensive
review exists in this context. This study conducts a systematic review to examine the potential directions
underlying the use of AI-enabled resources to assist learners in self-directed learning. The review targeted
articles from high-profile databases such as Web of Science and ScienceDirect. A range of AI resources,
including intelligent tutoring systems and conversational agents, were identified in the reviewed articles
for their roles in providing personalized learning support, question-answering assistance, timely feedback,
action recommendations, and interactive learning experiences that foster self-directed learning aspirations.
The review also demonstrated that AI-supported self-directed learning can have a positive effect on the
satisfaction of learners’ autonomy and competence needs. Besides that, prior knowledge of the learners
acts as a facilitator for the continued use of AI resources, thereby positively affecting the outcomes of self-
directed learning. Although AI in education is still a rare research field, future studies should focus, for
example, on the use of concrete AI tools, such as intelligent tutoring systems or dialog systems supporting
other forms of self-directed learning.

INDEX TERMS Artificial intelligence, dialogue systems, intelligent tutoring, self-paced learning.

I. INTRODUCTION capabilities have resulted in expanded applications of AI

AI-enhanced tools are increasingly being applied to a wide within educational contexts to facilitate learning [9], [10],
range of learning contexts. AI uses intelligent algorithms [11]. The application of AI tools for improving learning
that can support learning, decision-making, and even human- at diverse contexts is gaining increased interest among
like responses [1], [2]. Learning has become more flexible learners themselves [12], [13], [14]. Amid the rapid tech-
and effective with AI approaches such as natural language nological surge, novel learner-centered approaches, such as
processing, deep learning, and machine learning [3], [4]. For self-directed learning, have emerged and are now intertwined
example, artificial intelligence has streamlined and improved with technology.
the interactivity of learning, processing of information, deliv- Self-directed learning gained significant momentum dur-
ery of content, and acquisition of knowledge via intelligent ing the COVID-19 pandemic [15], [16], [17]. Rooted in
tutoring systems, dialogue agents, voice assistants, virtual learner autonomy, it empowers learners to navigate and dic-
reality, and gamification [5], [6], [7], [8]. These distinct tate the pace of their learning engagement. Furthermore,
self-directed learning enables continuous learning beyond
The associate editor coordinating the review of this manuscript and
the confines of the classroom and the direct influence of
approving it for publication was James Harland. a teacher [18], [19]. Technologies, particularly those with

2025 The Authors. This work is licensed under a Creative Commons Attribution 4.0 License.
VOLUME 13, 2025 For more information, see https://creativecommons.org/licenses/by/4.0/ 38387
 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

AI capabilities, have been considered viable support systems context-based sample sentences and exercise writing
for self-directed learning. Consequently, numerous studies skills [38]. Other research has explored AI-powered tools
have explored the design and implementation of AI-enabled that are created to provide real-time feedback and sugges-
tools to support self-directed learning. However, a systematic tions within self-study exercises, which gained popularity
review in this area remains lacking, prompting the current as a result of the pandemic [39], [40]. [41] designed the
study to provide a comprehensive overview of AI’s role in Personalized Self-Directed Learning Recommendation Sys-
self-directed learning. tem (PSDLR) and rendered it effective in order to equip
self-directed learners with proper suggestions and coun-
A. CONCEPTUALIZING SELF-DIRECTED LEARNING seling on the accessible e-learning or web-based content
Self-directed learning (SDL) is a strategy in which learn- alternatives. Further, if technology allows students to access
ers assume personal responsibility and guide the process of necessary content easily, they are more likely to engage
creating knowledge and achieving learning objectives. Using in self-directed learning [42]. It has been found through
available resources, learners practicing self-directed learning research that interaction with an AI-powered utility agent
independently identify their learning needs, set goals, and reduces cognitive load in self-directed learning systems [43].
explore relevant tasks [20], [21], [22]. Drawing from Gar- AI is evidently revolutionizing self-directed learning, and
rison’s perspective, it is evident that self-directed learners other AI utilities are being developed and introduced.
take initiative and accountability for their learning paths,
C. STUDY MODEL
as they possess a deeper understanding of their learning
requirements [23], [24]. Specifically, three essential elements Following is the study model (Figure 1) that represents the
self-monitoring, self-management, and motivation are central elements and characteristics of AI-supported self-directed
learning. It was proposed based on the research questions
to Garrison’s SDL model [25]. Motivation for learning and
and findings associated with them. The model summarizes
readiness to engage meaningfully in learning appropriate to
the key data constituting the overall architecture and findings
individualized goals are critical for effective self-directed
of the review. For instance, the left half of the model high-
learning, as noted by [22], [26], [27], and [28]. This model
lights two categories of AI-powered tools identified in the
of self-directed learning can be extended over long periods,
studies—Intelligent Tutoring Systems (ITS) and Dialog Sys-
where students engage in deep ways at a pace determined by
tems (DS)—that provide functions to support self-directed
their own interests.
learning. The other half of the model presents influencing
There is agreement on the general view that SDL and SRL
factors (personal autonomy, competence, prior knowledge)
share some similarities in that both concepts relate to facets of
and outcomes linked to AI-assisted self-directed learning.
learners’ engagement in self-directed learning, as evidenced
by the works of [29], [30], and [31]. In both approaches,
II. SYSTEMATIC REVIEW METHOD
the learner’s autonomy of their learning processes in SDL
The study methods adhere to the key requirements of a
and regulation of mechanisms in SRL are central aspects.
systematic review. Thus, pertinent search, screening, and
While theoretical distinctions exist between the two, their
selection protocols were followed.
implementation often overlaps [32].
A. SEARCH PROCEDURE
B. SELF-DIRECTED LEARNING THROUGH AI RESOURCES The study period spans from January 2020 to June 2023,
AI technologies are being utilized in learning to enable reflecting the heightened interest in self-directed learning
self-directed learning. Recent research has shown that AI during the COVID-19 pandemic. Although studies on AI
resources are revolutionizing the way self-directed learning in learning existed before this period, very few specifically
is negotiated and attained. Self-directed learning has been focused on self-directed learning. Moreover, articles pub-
made more accessible and interesting because of the pres- lished before 2020 are not included in the selected databases.
ence of AI resources. Intelligent applications and systems The aim was to catalog high-quality studies on implementing
are present and accessible on mobile devices [33], [34], [35]. AI-enabled tools to advance self-directed learning practices.
For instance, [36] assessed an AI-based fuzzy mechanism Consistent with the researchers’ goal of using high-
embedded in an intelligent tutoring system and validated quality papers, the initial search was limited to the Web
that it facilitates personalized learning through improved new of Science database. The search query used the following
knowledge gain, learner satisfaction, and motivation. In the keywords: ‘‘artificial intelligence’’ OR ‘‘AI’’ OR ‘‘conver-
same vein, [37] designed and assessed a tutoring app that sational agents’’ OR ‘‘intelligent tutoring systems’’ AND
enhanced learners’ statistical analysis competencies via self- ‘‘assisted’’ OR ‘‘supported’’ AND ‘‘self-directed learning’’
learning. OR ‘‘self-regulated learning.’’ To focus the search, it was
Aside from intelligent tutoring systems, AI-based recogni- narrowed to articles within the intended review period
tion technologies have also been shown to be very effective (2020 – 2023). The search yielded 259 related articles.
in self-learning. For example, created and applied the Due to the limited number of studies generated and
Smart Roam Lingo app to allow users to input substantial, the need for additional articles, further searches were

38388 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

FIGURE 1. AI-assisted SDL model.

conducted on the Elsevier-ScienceDirect database, yielding 1) VALIDATING THE SELECTION PROCESS

an additional 184 articles for screening. The input of search To ensure validity and minimize bias in the selection process,
keywords varied based on each database’s unique character- a systematic review protocol was developed, detailing the
istics. Researchers further narrowed the search to articles in search databases, search keywords, search limits, screening
the subject areas of education/computers, retrieving abstracts procedures, and the expected data to be extracted. Addition-
for further screening. ally, in Table 2, the year-wise selected articles were further
categorized to streamline management and analysis. This
B. CRITERIA DETERMINING INCLUSION AND EXCLUSION categorization involved assigning labels to each study, indi-
OF ARTICLES cating the categories of AI tools (e.g., Intelligent Tutoring
The identified articles were subjected to further screening Systems [ITS], Dialogue Systems [DS]) as well as their
based on the following predetermined inclusion and exclu- applied functions (e.g., tutoring, feedback, recommendation,
sion criteria (table.1). conversation, speech recognition, content generation).

C. SELECTION OF ARTICLES III. FINDINGS

The initial consideration for article selection was guided Consistent with the main objectives of the study, this review
by the inclusion and exclusion criteria established before- identifies the diverse roles of AI resources and their perfor-
hand. Strict selection procedures were followed to minimize mance in assisting self-directed learning, as well as explores
bias throughout the process. Explicit protocols were adhered limitations from previous studies. Based on the range of
to, ensuring that personal research interests did not influ- AI resources identified in the reviewed studies, intelligent
ence decisions. In cases where conflicting interests arose, tutoring systems and dialog systems are the two primary
the researchers consulted one another to reach a consensus, categories of AI systems that learners engage with for
allowing acceptable standards to prevail. As a result, the self-directed learning. While intelligent tutoring systems pro-
following procedures were implemented, establishing a clear vide learners with more adaptive and intelligent instruction
and systematic selection path. or feedback [44], dialog systems offer interactive learning
A total of 443 records were initially identified. However, support in real time [45]. Therefore, the findings capture
the majority focused either solely on AI or self-directed learn- the dimensions of AI resources and functions that support
ing (SDL), without addressing their intersection. During the self-directed learning practices, as well as the influencing
initial screening, 11 records were removed due to duplication, factors and learning outcomes underpinning students’ self-
and 416 were excluded based on the established exclu- learning.
sion criteria. Subsequently, further screening was conducted,
beginning with a close examination of titles, keywords, and A. AI-ENABLED FUNCTIONS AND RESOURCES THAT
abstracts. This was followed by a more detailed review of the ASSIST STUDENTS’ SELF-DIRECTED LEARNING
full text. Ultimately, only a minimal number of 16 articles met Drawing on the specific AI resources utilized in the reviewed
the inclusion criteria and were selected for review (Figure 2). studies, Figure 3 illustrates the key AI-enabled functions

VOLUME 13, 2025 38389

 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

TABLE 1. List of inclusion and exclusion criteria.

(n=4) that assist self-directed learning activities, including Similarly, [37] demonstrated that SDL-ITA is an effi-
personalized on-demand tutoring, query response, real-time cient provider of domain-specific content tutoring in a
feedback and recommendations, and interactive dialogue. self-directed learning context. Additionally, MENTOR sup-
Similarly, a series of AI-enabled tools (n=15) are uniquely ports peer tutoring in self-directed learning practices [47].
aligned with each of the AI-enabled functions identified, Meta Tutor (MT-ITS) was proven effective in facilitating
facilitating self-directed learning tasks. Two studies coin- students’ self-regulated learning [48], whereas BioWorld –
cided in their application of an intelligent personal assistant Intelligent Tutoring System (BW-ITS) is proven to effectively
(IPA) to assist learners’ self-directed learning. support clinical reasoning skills learning [49].

1) AI-ASSISTED PERSONALIZED TUTORING 2) AI-ASSISTED RESPONSE QUERY

A significant number of the reviewed studies (n=5) under- Findings of the review showed that one of the studies was
score the role of AI in providing personalized tutoring in on the application of an AI-supported tool to offer query
self-directed learning contexts. A variety of AI resources have response services. The query response task carried out by
been used to support self-directed learning through intelligent AI in self-study learning supports students’ inquiries through
instruction, including: instant answers. Intelligent Personal Assistant (IPA) or virtual
A review of the selected articles in Figure 4, showed a assistant is one of the AI technologies employed to optimize
positive correlation between the use of AI resources and students’ independent learning [51, 12.50]. The AI program
improvements in self-directed learning through personalized employs machine learning algorithms in comprehending and
tutoring. The AI-powered fuzzy mechanism operates within reacting to students’ inquiries or commands during indepen-
an intelligent tutoring system to enhance individualized dent learning tasks. The IPA intervention allows the students
learning. In their study, [46] applied the fuzzy mechanism to get sufficient support, leading to changes in their behavior
to identify learners’ needs, knowledge levels, and make due to enhanced self-regulatory capacities [50].
appropriate adjustments to the process of engaging in indi-
vidualized learning. This Fuzzy-Based Intelligent Tutoring 3) AI-ASSISTED REAL-TIME FEEDBACK AND
System (FB-ITS) provides learners with tutoring assistance in RECOMMENDATIONS
computer programming and has proven useful for increasing The pedagogy of self-learning has been heavily sup-
learner satisfaction and delivering new knowledge. ported by artificial intelligence through timely feedback and

38390 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

FIGURE 2. Flowchart illustrating the article selection procedure.

suggestions to learners. Timely feedback is quite essential personalized feedback according to their original
in learning since it enables one to gauge performance or writing [38].
progress [52]. In self-learning context, AI can act as an inter- The Fuzzy Expert System applies artificial intelligence
mediary in communicating to the learners how they grasp the methodology founded on fuzzy logic for information process-
course content. Certain AI tools (n=6) have been employed in ing and enhancing the learning process [53]. Integration of AI
the research covered herein to provide constructive feedback has enabled the provision of on-the-spot content recommen-
to support students’ independent learning processes. These dation to be very simple. A case in point is the Personalized
are presented in figure 5. Self-Directed Learning Recommendation System (PSDLR),
UBA is an adaptive model used in gamification to which is an essential AI system to support self-directed
offer complete and personalized learning experience. Ref- learning activities [41]. PSDLR is a recommendation system
erence [43] used UBA in their research to allow learners based on artificial intelligence that facilitates the provision of
to get instant and personalized feedback, along with instant recommendations on resources and gives suggestions
recommendations for cognitive activities. Likewise, the on potential alternatives while accessing e-learning or online
SRL writing tool incorporates AI recognition technolo- content during independent learning activities.
gies (AI-Sample Sentences and AI-Writing Feedback) In an attempt to enhance learning outcomes, [54] employed
that enhance the writing ability of students by offering the AI-enhanced tool known as Dashboard (DB) in their

VOLUME 13, 2025 38391

 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

TABLE 2. Influencing factors and the outcomes relating to AI-assisted SDL.

38392 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

TABLE 2. (Continued.) Influencing factors and the outcomes relating to AI-assisted SDL.

FIGURE 3. AI functions and related resources supporting self-directed learning.

VOLUME 13, 2025 38393

 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

FIGURE 4. Personalized tutoring enabled by AI.

research to assist students in their own learning. Through incentive to interact and supports knowledge consolidation
their research, they determined that the tool is significant in in self-regulated learning contexts. Additionally, the study
enhancing academic performance as well as self-regulatory by [58] showed that the use of the Dialog flow Chatbot (DC)
skills among students. Along the same lines, the Rule-Based serves as support for improving students’ learning outcomes
AI System (RBAIS) is employed to aid students’ self-study during self-regulated learning.
by giving instant feedback and is effective in promoting
learning performance among students [55]. B. INFLUENCING FACTORS AND LEARNING OUTCOMES
ARISING FROM AI-ASSISTED SELF-DIRECTED LEARNING
4) AI-ASSISTED INTERACTIVE DIALOGUE In addressing, Figure 7 below provides a list of the studies
These AI resources are applied to assist students’ self- and their related AI tools, along with details highlighting the
directed learning by facilitating interactive dialogue that indicative influencing factors and self-learning outcomes.
culminates in knowledge construction. The AI tools used for
interactive dialogue fall within the domain of conversational 1) SELF-LEARNING OUTCOMES AND THEIR INFLUENCING
agents and are valuable in providing students with a platform FACTORS
to engage in active learning characterized by high-level inter- The findings revealed a range of implied influencing factors
activity. Examples included in figure 6: that stimulate learners’ use of AI tools in supporting their
SIA is reported for its usefulness in supporting learn- self-directed learning, as indicated in Table 3 above. The
ers’ interactions during self-directed learning [56]. It is a reviewed articles showed that personal autonomy(n=7),
conversational agent that facilitates both written and oral competence(n=6), and previous knowledge(n=3) are the
language learning and can be accessed on mobile devices. dominant factors that serve as incentives to influence learn-
Similarly, [57] applied the AI Conversational Chatbot (AICC) ers’ sustained use of AI-enabled resources for self-directed
to determine the association between students’ prior knowl- learning in figure 8.
edge and self-regulated learning. Their study revealed that Based on Self-Determination Theory (SDT), both per-
the application of the AI chatbot provides learners with an sonal autonomy and competence are drivers of learning

38394 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

FIGURE 5. AI resources in self-directed learning.

FIGURE 6. AI tools for interactive learning.

satisfaction [57]. Personal autonomy provides self-directed of their learning. Competence, on the other hand, con-
learners with the freedom to decide on the path and pace notes learners’ feelings of effectiveness in utilizing the tools

VOLUME 13, 2025 38395

 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

FIGURE 7. AI – enabled resources.

FIGURE 8. Learning engagement factors.

available for self-directed learning. Similarly, the stud- AI tools to assist their self-directed learning tasks [53], [57],
ies showed that prior knowledge serves as a stimu- [58]. Notably, these influencing factors serve as catalysts
lant that enables learners to sustain engagement with for self-directed learning based on the unique capabilities

38396 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

of the AI tools adopted to assist in the learning occupied by students in self-learning via the application of AI
process. tools [59]. Additionally, the evidence from the studies under
The potential that AI-supported self-directed learning can review indicates that self-directed learning through AI helps
produce intended learning outcomes. As seen in figure 9, in achieving more learner engagement [36].
the research that was analyzed showed that the utilization of
AI-supported tools to facilitate tasks related to self-directed IV. DISCUSSION
learning allowed learners to achieve different outcomes like As the current research involves a review of studies on
skill acquisition, knowledge creation, cognitive development, AI-supported self-directed learning, trends relevant to AI
and fulfillment of needs, among others. functions and determinants supporting self-directed learning
Apart from this, the findings included varied method- have been determined. The reviewed studies provide insights
ologies through which the factors influence the product of into how different types of AI-supported tools can facilitate
AI-supported independent learning. Reference [43] examined self-directed learning activities and enable the accomplish-
the possibility of AI-assisted tools to better support learn- ment of desired learning outcomes. The above research
ers’ overall performance compared to independent learning highlights the importance of autonomy, interactivity, and
with a methodology known as ‘‘AI-induced guidance.’’ An relatedness as basic conditions that enable the thriving of
experiment employing an AI story game was applied in AI-supported self-directed learning. Extending these conclu-
conducting the study to avoid learning bottlenecks due to sions, this discourse provides an integrated analysis of the
insufficient cognitive load. The experiment demonstrated multi-dimensional support that self-directed learners can gain
that self-paced learning experimental group learners could from using AI-powered resources.
enhance story-learning performance with lower cognitive
load and continued attention. The experiment also high- A. THE SUPPORTIVE ROLE OF AI IN SELF-DIRECTED
lighted that learner autonomy in applying the tool influenced LEARNING
improvement in their learning performance. The review showed in figure 10 that AI tools are relevant
In a similar study, [12] tested the effectiveness of virtual in aiding self-learning through the provision of personal-
artificial intelligence agents like Amazon’s Alexa in aug- ized learning support, question-answering support, timely
menting learners’ overall performance in second-language feedback, action recommendations, and interactive learning
learning settings. The results showed that the learners felt support. Evidence from the reviewed articles suggests that AI
the use of Alexa was useful to their independent Japanese resources help learners gain on-demand tutoring, sustained
learning, which was enhanced through greater interaction and interactivity, increased content access, enhanced learning
engagement. The role of autonomy in such learning outcomes satisfaction, and reduced cognitive load while engaged in
is highlighted by the learners’ own statements as they applied self-directed learning. However, [49] showed inconsistencies
the tool beyond formal classroom contexts. in the effect that the use of an intelligent tutoring system
In addition, [37] study highlighted the promise of AI-based (BioWorld) had on medical students’ cognitive load.
technologies in facilitating improved self-directed learning
achievements in specific contexts, such as learning statistics. 1) PERSONALIZED LEARNING SUPPORT
A pre-test and post-test method was used by the researcher A significant number of the studies in this review reflect
that entailed the use of a tutoring app that was programmed on the application of AI resources to support personalized
with self-directed learning capabilities, wherein it was shown learning tasks. Typically, students engaged in self-directed
that students scored better on the post-test compared to learning become more reliant on opportunities that allow
the pre-test. Technology-enhanced student self-study enabled them to negotiate learning in a personalized manner. This
improvement in students’ statistical analysis competency makes the learning experience more engaging and meaning-
through feelings of autonomy when operating the technology. ful. Fortunately, the existence of AI-enabled tools provides
Research by [46] indicated that students are more moti- the nexus for seamless self-directed learning based on person-
vated while studying with AI-supported technologies such alized tutoring. Through the personalized tutoring function,
as intelligent tutoring systems during their own self-study. studies found that AI assists learners in their self-learning to
Students loved the interactive nature of this technology, as it gain and master new knowledge, improve analytical skills,
provided them with instant feedback about their learning and and scaffold learning [36], [48], [60].
highlighted areas in which they required further assistance. The personalized tutoring function is typically integrated
For their part, [55] study found no effect on learning out- into intelligent tutoring systems that can be used to assist
comes through the application of the Rule-Based AI System in personalized learning at any place and time. Additionally,
(RBAIS). The study, however, did find a significant enhance- the personalized tutoring support provided by these AI-driven
ment of self-regulated learning activity. tools mediates the self-learning process to fit the specific
Generally, motivation and engagement were the medi- needs of learners. This elevates personalized tutoring as one
ators of primary concern that affect the self-learning of the most cherished functions that AI tools provide in
results triggered by the application of AI tools by assisting self-learning. The reviewed studies indicated that
students. Such engagement concerns the quality of time AI resources are most suitable and promising in overcoming

VOLUME 13, 2025 38397

 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

FIGURE 9. AI tools in education.

FIGURE 10. AI tools in self-directed learning.

challenges associated with personalized tutoring and adaptive receive support to address gaps in their knowledge as well
learning. as needed assistance. Seemingly, this query function that AI
However, not all studies reported a high impact derived provides has not been given much application in self-learning
from the use of AI for personalized tutoring. The application activities. Perhaps conducting more studies in this regard
of the MENTOR AI tool produced contradictory results, thus could provide awareness of the issues and directions for the
warranting a need for the system to be improved [47]. application of virtual assistants in self-directed learning and
other learning contexts.
2) LEARNERS’ QUESTION-ANSWERING SUPPORT
Based on the findings presented in Figure 4, only a sin- 3) LEARNING SUPPORT THROUGH FEEDBACK AND ACTION
gle study accounts for the use of AI to provide learners RECOMMENDATIONS
with needed assistance in answering their questions during Based on the study findings presented in Figure 4, most
self-learning [50]. The AI-enabled response query function of the reviewed articles focus on the use of AI to support
assists learners in receiving content support through virtual students with timely feedback and recommendations in their
assistants. With this AI capability, learners are allowed to self-learning activities. This underscores the popular use

38398 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

of the AI-enabled feedback and recommendation functions. view may hold some merit, the valuable role of conversational
The real-time feedback and recommendation feature of AI chatbots in assisting self-learning cannot be ignored.
resources enable learners to gauge their learning performance Conclusively, this systematic review provides evidence of
and access appropriate subject content. A recent study proved the growing interest in and application of AI resources in
that the provision of effective feedback and recommendations assisting self-learning tasks. Based on their unique functions,
can support students’ overall self-learning [61], [62]. Such AI tools bear the capacity to transform learning and positively
is the benefit that AI resources provide for learners in their affect learners’ performance.
self-learning aspirations and engagements.
Most of the studies [38], [43], [53] reported consistent B. AI SUPPORT AND ITS IMPLICATIONS FOR
results that underpin the effectiveness of AI resources in SELF-DIRECTED LEARNING
enabling learners to derive satisfaction and improve their
Several of the reviewed studies explored the potential of
self-directed learning experience. For example, it has been
AI-enabled tools to improve learners’ overall outcomes when
demonstrated that students with a higher readiness for inde-
engaged in self-directed learning [12]. Consistent with the
pendent learning will have better and productive learning
main thrust in Figure 11, the emerging evidence from the
results [53]. Arguably, artificial intelligence significantly
review suggests that AI support for learners’ self-learning
enhances the learning impact more than other tools because
induces need satisfaction, resulting in desired learning out-
the learning process is rendered much more flexible, effec-
comes. Similarly, the role of learners’ prior knowledge in
tive, and efficient by providing essential feedback, support,
AI-assisted self-learning is recognizable. However, learners’
and guidance [45].
interactive competence comes into play in effective AI-
Besides, when students are involved in self-directed or self-
assisted self-directed learning.
regulated learning activities, the experience is improved by
the availability of tools that facilitate real-time feedback and
suggestions. Conversely, the absence of an effective mecha- 1) INFLUENCE ON LEARNERS’ NEED SATISFACTION
nism that can provide prompt feedback and recommendations The study findings indicated that need satisfaction plays
makes the self-learning process somewhat cumbersome for a pivotal role in sustaining learners’ use of AI resources
learners [54]. for self-directed learning tasks. Autonomy and competence
However, limitations exist regarding the level of effect are the need satisfaction prerequisites as theorized in the
induced by some of the AI resources applied in the studies. Self-Determination Theory [57]. Autonomy is a motivational
For example, while the study by [55] showed that AI-enabled element that compels students to embrace continued use of
tools can be used to provide personalized scaffolds in self- AI resources for self-study.
learning situations, it found little effect on learning outcomes. In line with the findings presented, most of the studies
Perhaps this drawback is attributable to the fact that some indicated that factors related to personal autonomy played a
of the AI resources utilized to support self-directed learn- significant role in students continued use of AI tools designed
ing are not specifically designed for application in learning to facilitate their autonomous learning. In general, students
contexts [6]. preferred to demonstrate better self-learning performances,
as reflected in higher levels of motivation and engagement.
For example, learners in [43] study had better learning
4) INTERACTIVE LEARNING SUPPORT outcomes, manifested as greater performance, because of
A more functional attribute ascribed to AI tools is interactiv- interacting with the Utility-Based Agent (UBA).
ity. This interactive dialogue function is a common attribute Artificial Intelligence resources have a significant effect
of conversational agents such as conversational chatbots. on the satisfaction of learners’ competence needs. This is
As shown in Figure 2, three AI-enabled tools emerged from traced to the Self-Determination Theory, which is concerned
the studies reviewed [56.57]. These conversational agents with the conditions for satisfaction. Competence is traced
depend on AI entities such as machine learning and natural to the perception by the learners of their effectiveness in
language processing to perform related tasks. The use of utilizing AI in facilitating their autonomous learning pro-
conversational agents can assist in promoting self-directed cesses. Relatedly, competence connects with learners’ beliefs
learning by delivering interactive learning support. about usability and satisfaction with the AI tools used for
The studies showed that the related AI resources are cat- learning [43].
alysts for learning through interactivity. As conversational Several studies [36], [47], [48], [49], [54], [56] under-
agents accelerate interactive dialogue, they can be valuable scored the influencing role of learners’ competence on their
tools that stimulate learning, induce engagement, and propel AI-assisted self-learning outcomes. The influencing factor
learning outcomes [58]. In essence, they can be valuable tools relating to competence gives rise to more diversified outcome
for learning. indicators, including skill development, improved learn-
However, [57] argued that, although conversational chat- ing, and enhanced cognitive ability. Arguably, competence
bots are effective for knowledge consolidation, they are not seems to be the most compelling factor influencing learners’
quite suitable for self-regulated learning tasks. While their AI-assisted self-directed learning outcomes.

VOLUME 13, 2025 38399

 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

FIGURE 11. AI in self-directed learning.

2) EFFECTS OF LEARNERS’ PRIOR KNOWLEDGE scaffolding learning. In this regard, how students engage with
Prior knowledge is a notable influencer of students’ achieve- AI agents in independent learning is largely important in
ments and learning outcomes [63]. Surprisingly, only a few generating the intended result [66], [67], [68], [69], [70], [71]
of the studies [53], [57], [58] reflect on prior knowledge Overall, the reviewed literature findings have revealed the
as a catalyst for AI-assisted self-learning outcomes. Signif- various ways in which students’ self-directed learning can be
icantly, these few studies provide strong support to show that supported by using AI resources. The various AI resources
prior knowledge sustains learners’ engagement with AI in and their respective capabilities simply present overwhelm-
assisting self-learning and hence fosters the improvement of ing evidence that reveals how much impact these emerging
performance. technologies are likely to have on education and pedagogy.
The studies established a strong association between As more students strive to undertake self-directed learning
prior knowledge and improvement in learning outcomes. activities, using AI resources becomes more essential.
For instance, it concluded that prior knowledge influences
learning outcomes in an AI-assisted self-learning context. V. CONCLUSION
However, [57] study did not directly compare the effective- This review paper explains the pivotal role of artificial
ness of AI-enabled tools against other types of interventions intelligence resources in facilitating different aspects of self-
or teaching approaches in improving learners’ overall out- directed learning. Although the concept of self-directed
comes related to self-learning. Nevertheless, it provides learning (SDL) has been implemented in learning settings
fundamental insights into how AI-enabled technologies may for a considerable number of decades, its dynamic expres-
offer unique benefits for supporting individualized and adap- sions in informal and formal learning contexts continue to
tive learning experiences. unfold. Essentially, an exploration of AI functions, influential
factors, and limitations in assisting self-directed learning has
3) LEARNERS’ INTERACTIVE COMPETENCE provided additional scholarly perspectives on the growing
The success of AI-supported self-learning can be influenced influence that AI exerts in facilitating students’ learning activ-
by learners’ interactive competence. This is the capacity ities. Notably, the identified AI resources are presented as
of learners to engage in productive interactions with their facilitators of intelligent functions that suitably assist stu-
environment—i.e., AI—to achieve learning objectives [64], dents’ self-directed learning aspirations. The results indicate
[65]. Although the literature review offers significant evi- that AI resources perform supportive roles in self-learning
dence indicating the positive impact of AI on self-learning, by providing learners with personalized learning support,
it is important to acknowledge the potential issues that may question-answering support, learning support through timely
emanate from learners’ interactions. Interactive competency feedback and action recommendations, and interactive sup-
is undoubtedly necessary to enable students to actualize the port. Moreover, the use of AI has a strong influence on
pedagogical gains of AI’s personalized learning assistance, learners’ self-learning practices and outcomes, encompassing
question-answering function, prompt feedback and hints, and skills acquisition and knowledge building across disciplines.

38400 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

The review shows that learners’ needs satisfaction and prior [11] O. Tapalova and N. Zhiyenbayeva, ‘‘Artificial intelligence in education:
knowledge serve as catalysts for improved performance and AIEd for personalised learning pathways,’’ Electron. J. E-Learn., vol. 20,
no. 5, pp. 639–653, Dec. 2022, doi: 10.34190/ejel.20.5.2597.
skills competence in AI-assisted SDL environments. [12] G. Dizon, D. Tang, and Y. Yamamoto, ‘‘A case study of using
Alexa for out-of-class, self-directed Japanese language learning,’’ Com-
VI. LIMITATIONS put. Educ., Artif. Intell., vol. 3, Oct. 2022, Art. no. 100088, doi:
10.1016/j.caeai.2022.100088.
Notably, this review did not capture data from studies [13] A. Y. Q. Huang, O. H. T. Lu, and S. J. H. Yang, ‘‘Effects of arti-
employing a qualitative perspective that could provide a ficial intelligence–enabled personalized recommendations on learners’
deeper understanding of how self-directed learning is sup- learning engagement, motivation, and outcomes in a flipped class-
room,’’ Comput. Educ., vol. 194, Mar. 2023, Art. no. 104684, doi:
ported with AI. Moreover, the investigation was limited to 10.1016/j.compedu.2022.104684.
studies published between January 2020 and June 2023, [14] L. Markauskaite, R. Marrone, O. Poquet, S. Knight,
with search terms restricted to ‘‘artificial intelligence,’’ R. Martinez-Maldonado, sS. Howard, J. Tondeur, M. De Laat,
‘‘conversational agents,’’ and ‘‘intelligent tutoring systems.’’ S. Buckingham Shum, D. Gašević, and G. Siemens, ‘‘Rethinking
the entwinement between artificial intelligence and human learning:
Additionally, due to the paucity of studies on AI-assisted What capabilities do learners need for a world with AI?’’ Comput.
SDL, a more generalized approach was adopted, explor- Educ., Artif. Intell., vol. 3, Apr. 2022, Art. no. 100056, doi:
ing studies on both self-directed and self-regulated learning. 10.1016/j.caeai.2022.100056.
[15] M. C. Maphalala, R. G. Mkhasibe, and D. W. Mncube, ‘‘Online learning
Future studies could specifically explore the utilization of as a catalyst for self-directed learning in universities during the COVID-19
a specific category of AI tools, such as intelligent tutoring pandemic,’’ Res. Social Sci. Technol., vol. 6, no. 2, pp. 233–248, Sep. 2021,
systems or dialog systems, in assisting self-directed learning. doi: 10.46303/ressat.2021.25.
[16] W. Sun, J.-C. Hong, Y. Dong, Y. Huang, and Q. Fu, ‘‘Self-directed learning
Furthermore, the implications of learners’ interactive com-
predicts online learning engagement in higher education mediated by per-
petence in AI-assisted self-directed learning received limited ceived value of knowing learning goals,’’ Asia Pacific Educ. Researcher,
attention in this review. Given possible interactive compe- vol. 32, no. 3, pp. 307–316, Jun. 2023, doi: 10.1007/s40299-022-00653-6.
tence gaps, studies focusing on the ways learners interact with [17] U. Noor, M. Younas, H. S. Aldayel, R. Menhas, and X. Qingyu, ‘‘Learn-
ing behavior, digital platforms for learning and its impact on university
AI agents in self-learning situations could be carried out. student’s motivations and knowledge development,’’ Frontiers Psychol.,
vol. 13, Nov. 2022, Art. no. 933974, doi: 10.3389/fpsyg.2022.933974.
REFERENCES [18] E. Sumuer, ‘‘Factors related to college students’ self-directed learning with
technology,’’ Australas. J. Educ. Technol., vol. 34, no. 4, Sep. 2018, doi:
[1] T. Baker and L. Smith, ‘‘Educ-AI-tion rebooted? Exploring the future 10.14742/ajet.3142.
of artificial intelligence in schools and colleges,’’ Nesta Found.,
[19] M. Younas, U. Noor, X. Zhou, R. Menhas, and X. Qingyu, ‘‘COVID-19,
London, U.K., Tech. Rep., 2019. [Online]. Available: https://media.
students satisfaction about e-learning and academic achievement: Medi-
nesta.org.uk/documents/Future_of_AI_and_education
ating analysis of online influencing factors,’’ Frontiers Psychol., vol. 13,
[2] K.-Y. Tang, C.-Y. Chang, and G.-J. Hwang, ‘‘Trends in artificial Aug. 2022, doi: 10.3389/fpsyg.2022.948061.
intelligence-supported e-learning: A systematic review and co-citation
[20] T. H. Morris, G. Koutsouris, L. Stentiford, and N. Bremner, ‘‘Self-directed
network analysis (1998–2019),’’ Interact. Learn. Environments, vol. 31,
learning—A framework for inclusion ‘In’ and ‘Through’ education—A
no. 4, pp. 2134–2152, May 2023.
systematic review,’’ Rev. Educ., vol. 13, no. 1, Apr. 2025, Art. no. e70028.
[3] S.-T. Chu, G.-J. Hwang, and Y.-F. Tu, ‘‘Artificial intelligence-based
robots in education: A systematic review of selected SSCI publications,’’ [21] M. Zhu, S. Berri, R. Koda, and Y.-J. Wu, ‘‘Exploring students’ self-directed
Comput. Educ., Artif. Intell., vol. 3, May 2022, Art. no. 100091, doi: learning strategies and satisfaction in online learning,’’ Educ. Inf. Tech-
10.1016/j.caeai.2022.100091. nol., vol. 29, no. 3, pp. 2787–2803, Feb. 2024, doi: 10.1007/s10639-023-
11914-2.
[4] X. Wang, M. Younas, Y. Jiang, M. Imran, and N. Almusharraf,
‘‘Transforming education through blockchain: A systematic review [22] T. H. Morris, ‘‘Self-directed learning: A fundamental competence in a
of applications, projects, and challenges,’’ IEEE Access, vol. 13, rapidly changing world,’’ Int. Rev. Educ., vol. 65, no. 4, pp. 633–653,
pp. 13264–13284, 2025. [Online]. Available: https://ieeexplore. Aug. 2019, doi: 10.1007/s11159-019-09793-2.
ieee.org/document/10804766 [23] M. Zhu, C. J. Bonk, and M. Y. Doo, ‘‘Self-directed learning in
[5] J. H. Al Shamsi, M. Al-Emran, and K. Shaalan, ‘‘Understanding key MOOCs: Exploring the relationships among motivation, self-monitoring,
drivers affecting students’ use of artificial intelligence-based voice assis- and self-management,’’ Educ. Technol. Res. Develop., vol. 68, no. 5,
tants,’’ Educ. Inf. Technol., vol. 27, no. 6, pp. 8071–8091, Jul. 2022, doi: pp. 2073–2093, Oct. 2020, doi: 10.1007/s11423-020-09747-8.
10.1007/s10639-022-10947-3. [24] M. Imran, N. Almusharraf, M. S. Abdellatif, and M. Y. Abbasova,
[6] H. Khosravi, S. B. Shum, G. Chen, C. Conati, Y.-S. Tsai, J. Kay, S. Knight, ‘‘Artificial intelligence in higher education: Enhancing learning
R. Martinez-Maldonado, S. Sadiq, and D. Gašević, ‘‘Explainable artificial systems and transforming educational paradigms,’’ Int. J. Interact.
intelligence in education,’’ Comput. Educ., Artif. Intell., vol. 3, May 2022, Mobile Technol. (iJIM), vol. 18, no. 18, pp. 34–48, Sep. 2024, doi:
Art. no. 100074, doi: 10.1016/j.caeai.2022.100074. 10.3991/ijim.v18i18.49143.
[7] V. Srinivasan, ‘‘AI & learning: A preferred future,’’ Comput. [25] D. R. Garrison, ‘‘Self-directed learning: Toward a comprehensive model,’’
Educ., Artif. Intell., vol. 3, Aug. 2022, Art. no. 100062, doi: Adult Educ. Quart., vol. 48, no. 1, pp. 18–33, Nov. 1997.
10.1016/j.caeai.2022.100062. [26] C. J. Bonk, M. M. Lee, X. Kou, S. Xu, and F. R. Sheu, ‘‘Understanding
[8] M. Afzaal, X. Shanshan, D. Yan, and M. Younas, ‘‘Mapping artifi- the self-directed online learning preferences, goals, achievements, and
cial intelligence integration in education: A decade of innovation and challenges of MIT open course ware subscribers,’’ J. Educ. Technol. Soc.,
impact (2013–2023)—A bibliometric analysis,’’ IEEE Access, vol. 12, vol. 18, no. 2, pp. 349–368, 2015.
pp. 113275–113299, 2024, doi: 10.1109/ACCESS.2024.3443313. [27] T. E. T. van Woezik, J. J.-J. Koksma, R. P. B. Reuzel, D. C. Jaarsma,
[9] A. Bhutoria, ‘‘Personalized education and artificial intelligence in the and G. J. van der Wilt, ‘‘There is more than ‘I’ in self-directed learn-
United States, China, and india: A systematic review using a human- ing: An exploration of self-directed learning in teams of undergraduate
in-the-loop model,’’ Comput. Educ., Artif. Intell., vol. 3, Sep. 2022, students,’’ Med. Teacher, vol. 43, no. 5, pp. 590–598, May 2021, doi:
Art. no. 100068, doi: 10.1016/j.caeai.2022.100068. 10.1080/0142159x.2021.1885637.
[10] S. Järvelä, A. Nguyen, and A. Hadwin, ‘‘Human and artificial intel- [28] O. T. Adigun, N. Mpofu, and M. C. Maphalala, ‘‘Fostering self-directed
ligence collaboration for socially shared regulation in learning,’’ Brit. learning in blended learning environments: A constructivist perspective
J. Educ. Technol., vol. 54, no. 5, pp. 1057–1076, Sep. 2023, doi: in higher education,’’ Higher Educ. Quart., vol. 79, no. 1, Jan. 2025,
10.1111/bjet.13325. Art. no. e12572.

VOLUME 13, 2025 38401

 M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

[29] J. Robertson, ‘‘The educational affordances of blogs for self-directed [48] R. Azevedo, F. Bouchet, M. Duffy, J. Harley, M. Taub, G. Trevors,
learning,’’ Comput. Educ., vol. 57, no. 2, pp. 1628–1644, Sep. 2011, doi: E. Cloude, D. Dever, M. Wiedbusch, F. Wortha, and R. Cerezo, ‘‘Lessons
10.1016/j.compedu.2011.03.003. learned and future directions of MetaTutor: Leveraging multichannel data
[30] K. Saks and Ä. Leijen, ‘‘Distinguishing self-directed and self- to scaffold self-regulated learning with an intelligent tutoring system,’’
regulated learning and measuring them in the e-learning context,’’ Frontiers Psychol., vol. 13, Jun. 2022, doi: 10.3389/fpsyg.2022.813632.
Proc.-Social Behav. Sci., vol. 112, pp. 190–198, Feb. 2014, doi: [49] T. Wang, S. Li, X. Huang, Z. Pan, and S. P. Lajoie, ‘‘Examining students’
10.1016/j.sbspro.2014.01.1155. cognitive load in the context of self-regulated learning with an intelli-
[31] P. J. Woods and Y. Copur-Gencturk, ‘‘Examining the role of student- gent tutoring system,’’ Educ. Inf. Technol., vol. 28, no. 5, pp. 5697–5715,
centered versus teacher-centered pedagogical approaches to self-directed May 2023, doi: 10.1007/s10639-022-11357-1.
learning through teaching,’’ Teach. Teacher Educ., vol. 138, Feb. 2024, [50] E. Pogorskiy and J. F. Beckmann, ‘‘From procrastination to engagement?
Art. no. 104415. An experimental exploration of the effects of an adaptive virtual assistant
[32] A. Voskamp, E. Kuiper, and M. Volman, ‘‘Teaching practices for self- on self-regulation in online learning,’’ Comput. Educ., Artif. Intell., vol. 4,
directed and self-regulated learning: Case studies in Dutch innovative Feb. 2023, Art. no. 100111, doi: 10.1016/j.caeai.2022.100111.
secondary schools,’’ Educ. Stud., vol. 48, no. 6, pp. 772–789, Nov. 2022, [51] M. M. Asad, A. Basit, P. Churi, and N. Almusharraf, ‘‘Lifelong learning
doi: 10.1080/03055698.2020.1814699. paradigm and socioeconomic sustainability: Towards employment growth
[33] K. Sriwisathiyakun and C. Dhamanitayakul, ‘‘Enhancing digital literacy of vocational trainers in Pakistan,’’ Educ. + Training, vol. 66, no. 9,
with an intelligent conversational agent for senior citizens in Thailand,’’ pp. 1262–1275, Nov. 2024, doi: 10.1108/et-11-2023-0461.
Educ. Inf. Technol., vol. 27, no. 5, pp. 6251–6271, Jun. 2022, doi: [52] L. Zheng, Y. Fan, B. Chen, Z. Huang, LeiGao, and M. Long, ‘‘An AI-
10.1007/s10639-021-10862-z. enabled feedback-feedforward approach to promoting online collaborative
[34] X. Lin, ‘‘Exploring the role of ChatGPT as a facilitator for motivating learning,’’ Educ. Inf. Technol., vol. 29, no. 9, pp. 11385–11406, Jun. 2024,
self-directed learning among adult learners,’’ Adult Learn., vol. 35, no. 3, doi: 10.1007/s10639-023-12292-5.
pp. 156–166, Aug. 2024. [53] D. B. Lasfeto and S. Ulfa, ‘‘Modeling of online learning strategies based
[35] M. Imran and N. Almusharraf, ‘‘Google Gemini as a next generation AI on fuzzy expert systems and self-directed learning readiness: The effect on
educational tool: A review of emerging educational technology,’’ Smart learning outcomes,’’ J. Educ. Comput. Res., vol. 60, no. 8, pp. 2081–2104,
Learn. Environ., vol. 11, no. 1, p. 22, May 2024, doi: 10.1186/s40561-024- Jan. 2023, doi: 10.1177/07356331221094249.
00310-z. [54] M. Afzaal, A. Zia, J. Nouri, and U. Fors, ‘‘Informative feedback
[36] K. Chrysafiadi, M. Virvou, G. A. Tsihrintzis, and I. Hatzilygeroudis, and explainable AI-based recommendations to support Students’ self-
‘‘Evaluating the user’s experience, adaptivity and learning outcomes of regulation,’’ Technol., Knowl. Learn., vol. 29, no. 1, pp. 331–354,
a fuzzy-based intelligent tutoring system for computer programming Mar. 2024, doi: 10.1007/s10758-023-09650-0.
for academic students in Greece,’’ Educ. Inf. Technol., vol. 28, no. 6,
[55] L. Lim, M. Bannert, J. van der Graaf, S. Singh, Y. Fan, S. Surendrannair,
pp. 6453–6483, Jun. 2023, doi: 10.1007/s10639-022-11444-3.
M. Rakovic, I. Molenaar, J. Moore, and D. Gašević, ‘‘Effects of real-
[37] S. Dangprasert, ‘‘Effects on using tutoring application in integration time analytics-based personalized scaffolds on students’ self-regulated
with self-directed learning to improve statistical analysis skills,’’ TEM J., learning,’’ Comput. Hum. Behav., vol. 139, Feb. 2023, Art. no. 107547, doi:
vol. 2021, pp. 63–68, Feb. 2021. 10.1016/j.chb.2022.107547.
[38] W.-Y. Hwang, R. Nurtantyana, S. W. D. Purba, U. Hariyanti, [56] J. Jeon, ‘‘Exploring a self-directed interactive app for informal EFL learn-
Y. Indrihapsari, and H. D. Surjono, ‘‘AI and recognition technologies ing: A self-determination theory perspective,’’ Educ. Inf. Technol., vol. 27,
to facilitate English as foreign language writing for supporting no. 4, pp. 5767–5787, May 2022, doi: 10.1007/s10639-021-10839-y.
personalization and contextualization in authentic contexts,’’ J.
[57] Q. Xia, T. K. F. Chiu, C. S. Chai, and K. Xie, ‘‘The mediating effects of
Educ. Comput. Res., vol. 61, no. 5, pp. 1008–1035, Sep. 2023, doi:
needs satisfaction on the relationships between prior knowledge and self-
10.1177/07356331221137253.
regulated learning through artificial intelligence chatbot,’’ Brit. J. Educ.
[39] G. Deeva, D. Bogdanova, E. Serral, M. Snoeck, and J. De Weerdt, ‘‘A
Technol., vol. 54, no. 4, pp. 967–986, Jul. 2023.
review of automated feedback systems for learners: Classification frame-
work, challenges and opportunities,’’ Comput. Educ., vol. 162, Mar. 2021, [58] M. C. Sáiz-Manzanares, R. Marticorena-Sánchez, L. J. Martín-Antón,
Art. no. 104094, doi: 10.1016/j.compedu.2020.104094. I. G. Díez, and L. Almeida, ‘‘Perceived satisfaction of university stu-
dents with the use of chatbots as a tool for self-regulated learn-
[40] E. Esiyok, S. Gokcearslan, and K. G. Kucukergin, ‘‘Acceptance of
ing,’’ Heliyon, vol. 9, no. 1, Jan. 2023, Art. no. e12843, doi:
educational use of AI chatbots in the context of self-directed learn-
10.1016/j.heliyon.2023.e12843.
ing with technology and ICT self-efficacy of undergraduate stu-
dents,’’ Int. J. Hum. Comput. Interact., vol. 41, no. 1, pp. 641–650, [59] Y. Wu, D. Rough, A. Bleakley, J. Edwards, O. Cooney, P. R. Doyle,
Jan. 2025. L. Clark, and B. R. Cowan, ‘‘See what I’m saying? Comparing intelligent
[41] T. B. Lalitha and P. S. Sreeja, ‘‘Personalised self-directed learning recom- personal assistant use for native and non-native language speakers,’’ 2020,
mendation system,’’ Proc. Comput. Sci., vol. 171, pp. 583–592, Sep. 2020, arXiv:2006.06328.
doi: 10.1016/j.procs.2020.04.063. [60] M. Younas, Y. Dong, R. Menhas, X. Li, Y. Wang, and U. Noor, ‘‘Alleviating
[42] S. Karimi, ‘‘Do learners’ characteristics matter? An exploration of the effects of the COVID-19 pandemic on the physical, psychological
mobile-learning adoption in self-directed learning,’’ Comput. Hum. health, and wellbeing of students: Coping behavior as a mediator,’’ Psy-
Behav., vol. 63, pp. 769–776, Oct. 2016, doi: 10.1016/j.chb.2016. chol. Res. Behav. Manage., vol. Volume 16, pp. 5255–5270, Dec. 2023,
06.014. doi: 10.2147/prbm.s441395.
[43] C. Ferguson, E. L. van den Broek, and H. van Oostendorp, ‘‘AI- [61] M. Afzaal, J. Nouri, A. Zia, P. Papapetrou, U. Fors, Y. Wu, X. Li, and
induced guidance: Preserving the optimal zone of proximal development,’’ R. Weegar, ‘‘Explainable AI for data-driven feedback and intelligent action
Comput. Educ., Artif. Intell., vol. 3, Apr. 2022, Art. no. 100089, doi: recommendations to support students self-regulation,’’ Frontiers Artif.
10.1016/j.caeai.2022.100089. Intell., vol. 4, Nov. 2021, Art. no. 723447, doi: 10.3389/frai.2021.723447.
[44] H. Guo, X. Zhuang, and T. Rabczuk, ‘‘A deep collocation method for the [62] M. Younas and Y. Dong, ‘‘The impact of using animated movies in learning
bending analysis of Kirchhoff plate,’’ 2021, arXiv:2102.02617. English language vocabulary: An empirical study of Lahore, Pakistan,’’
[45] L. Chen, P. Chen, and Z. Lin, ‘‘Artificial intelligence in educa- Sage Open, vol. 14, no. 2, Apr. 2024, doi: 10.1177/21582440241258398.
tion: A review,’’ IEEE Access, vol. 8, pp. 75264–75278, 2020, doi: [63] T. Hailikari, N. Katajavuori, and S. Lindblom-Ylanne, ‘‘The relevance of
10.1109/ACCESS.2020.2988510. prior knowledge in learning and instructional design,’’ Amer. J. Pharma-
[46] K. Chrysafiadi, S. Papadimitriou, and M. Virvou, ‘‘Cognitive-based adap- ceutical Educ., vol. 72, no. 5, p. 113, 2008.
tive scenarios in educational games using fuzzy reasoning,’’ Knowl.-Based [64] W.-T. Fu, J. Chin, W. Dong, and Q. V. Liao, ‘‘Interactive skills and dual
Syst., vol. 250, Aug. 2022, Art. no. 109111. learning processes,’’ in Encyclopedia of the Sciences of Learning. Boston,
[47] S. C. Tan, C. K. Looi, Y. L. Cheung, S. H. Chung, S. J. Lims, MA, USA: Springer, 2012, doi: 10.1007/978-1-4419-1428-6_1633.
and W. H. Wong, ‘‘Designing and evaluating a mobile peer tutor- [65] M. Younas, Y. Dong, G. Zhao, R. Menhas, L. Luan, and U. Noor, ‘‘Unveil-
ing application: A cultural historical activity theory approach,’’ Inter- ing digital transformation and teaching prowess in English education
act. Learn. Environ., vol. 31, no. 8, pp. 4806–4817, Nov. 2023, doi: during COVID-19 with structural equation modelling,’’ Eur. J. Educ.,
10.1080/10494820.2021.1983608. vol. 60, no. 1, Mar. 2025, Art. no. e12818, doi: 10.1111/ejed.12818.

38402 VOLUME 13, 2025

M. Younas et al.: Comprehensive Systematic Review of AI-Driven Approaches to Self-Directed Learning

[66] J. Abdullah, W. N. Mohd-Isa, and M. A. Samsudin, ‘‘Virtual reality DINA ABDEL SALAM EL-DAKHS is currently a Professor with the Linguis-
to improve group work skill and self-directed learning in problem- tics and Translation Department (LTD), College of Humanities and Sciences,
based learning narratives,’’ Virtual Reality, vol. 23, no. 4, pp. 461–471, Prince Sultan University, Riyadh, Saudi Arabia. She is also the LTD Chair
Dec. 2019, doi: 10.1007/s10055-019-00381-1. and the Leader of the Language and Communication Research Laboratory
[67] H. B. Essel, D. Vlachopoulos, A. Tachie-Menson, E. E. Johnson, and (LCRL). Her current research interests include second language acquisition
P. K. Baah, ‘‘The impact of a virtual teaching assistant (chatbot) on stu- (SLA), psycholinguistics, pragmatics, and discourse analysis.
dents’ learning in Ghanaian higher education,’’ Int. J. Educ. Technol.
Higher Educ., vol. 19, no. 1, pp. 1–19, Nov. 2022, doi: 10.1186/s41239-
022-00362-6.
[68] T. K. F. Chiu, Q. Xia, X. Zhou, C. S. Chai, and M. Cheng,
‘‘Systematic literature review on opportunities, challenges, and future
research recommendations of artificial intelligence in education,’’ Com- YICUN JIANG received the Ph.D. degree in
put. Education: Artif. Intell., vol. 4, Jan. 2023, Art. no. 100118, doi: English from Lingnan University, Hong Kong.
10.1016/j.caeai.2022.100118. He is currently an Associate Professor with the
[69] W. Toh and D. Kirschner, ‘‘Self-directed learning in video games, School of Foreign Languages, Shenzhen Tech-
affordances and pedagogical implications for teaching and learn- nology University (SZTU). Previously, he taught
ing,’’ Comput. Educ., vol. 154, Sep. 2020, Art. no. 103912, doi: with Shandong Technology and Business Univer-
10.1016/j.compedu.2020.103912. sity before coming to SZTU. He was a Special
[70] C. Zhai and S. Wibowo, ‘‘A systematic review on artificial intelligence Research Fellow with the Institute of Semiotics
dialogue systems for enhancing English as foreign language students’ and Media Studies (ISMS), Sichuan Univer-
interactional competence in the university,’’ Comput. Educ., Artif. Intell., sity. His research interests include semiotics and
vol. 4, Oct. 2023, Art. no. 100134, doi: 10.1016/j.caeai.2023.100134. metaphor study, theoretical linguistics, and cross-cultural study.
[71] B. J. Zimmerman, ‘‘Becoming a self-regulated learner: An overview,’’
Theory Into Pract., vol. 41, no. 2, pp. 64–70, May 2002, doi:
10.1207/s15430421tip4102_2.

MUHAMMAD YOUNAS received the Ph.D.

degree in higher education from Soochow Uni-
versity, China. He is currently a Researcher with
the Language and Communication Research Lab-
oratory, Prince Sultan University, Saudi Arabia.
Previously, he was a Research Fellow with the
School of Educational Technology, Beijing Nor-
mal University. He has published academic articles
in various SSCI and Scopus-indexed journals
related to language teacher education and techno-
logical teaching/learning. His research interests include artificial intelligence
in education, language teacher education, CALL, e-learning in higher
education, technology-enhanced learning, MALL, and virtual learning envi-
ronments (VLE). He received Jiangsu Province Outstanding International
Researcher Award from Soochow University.

VOLUME 13, 2025 38403