Student Perceptions of Generative Artificial Intelligence: Investigating Utilization, Benefits, and Challenges in Higher Education

Abstract

This research explores the use of Generative Artificial Intelligence (GenAI) tools among higher education students in Saudi Arabia, aiming to understand their current perceptions of these technologies. This study utilizes the Technology Acceptance Model (TAM) and the theory of Task-Technology Fit (TTF) to examine students’ utilization, perceived benefits, and challenges associated with these tools. A cross-sectional survey was conducted, yielding 859 responses. The findings indicate that 78.7% of students frequently use GenAI tools, while 21.3% do not, often due to a lack of knowledge or interest. ChatGPT emerged as the most widely used GenAI tool, utilized by 86.2% of respondents, followed by other tools like Gemini, Socratic, and CoPilot. Students primarily use these tools for defining or clarifying concepts, translation, generating ideas in writing, and summarizing academic literature. They cite benefits such as ease of access, time-saving, and instant feedback. However, they express concerns about the challenges, including subscription fees, unreliable information, plagiarism, reduced human-to-human interaction, and impacts on learning autonomy. This study underscores the need for increased awareness, ethical guidelines, and robust academic integrity measures to ensure the responsible use of GenAI tools in educational settings. These findings highlight the need for a balanced utilization of GenAI tools in higher education that maximizes benefits while addressing potential challenges and guides the development of policies, curricula, and support systems.

1. Introduction

Artificial Intelligence (AI) is gaining popularity in higher education as a growing field of research and application. AI is defined as “computing systems that are able to engage in human-like processes such as learning, adapting, synthesizing, self-correction and the use of data for complex processing tasks” [1] (p. 2). Within the broad domain of AI, Generative Artificial Intelligence (GenAI) represents a particularly promising technology that can create new and original content such as text, images, audio, and video using advanced language models and generative algorithms rather than just processing or analyzing existing data [2,3]. The literature reveals that AI has applications across diverse educational disciplines, with researchers from fields such as education, computer science, tourism, music, and public affairs exploring the potential of AI for educational purposes [4].

A recent study by [5] found that while university students in Hong Kong are optimistic about the integration of GenAI as part of the broader technological evolution, they also have reservations that educators and policymakers must carefully consider as they work to incorporate these advanced tools into the academic setting. AI policies have been developed and implemented globally, and the literature explores the specific policies and guidance related to the use of AI in education, including the latest UNESCO recommendations on the ethical deployment of AI in education and research [2]. UNESCO has issued recommendations on the ethical use of AI, outlining key principles such as maintaining safety and human oversight, ensuring transparency and fairness, respecting privacy and data protection, building awareness, and collaborating with stakeholders [6]. Understanding students’ perceptions and concerns regarding the integration of GenAI tools is crucial for facilitating their successful implementation in higher education to enhance teaching and learning.

As GenAI becomes increasingly integrated into the education sector, researchers have been motivated to explore the literature on its applications and impacts within educational environments [4]. Moreover, recent studies in higher education by [7,8] have highlighted the dearth of empirical evidence supporting claims regarding the advantages and challenges of AI in higher education, especially concerning GenAI. This lack of empirical research emphasizes the need for further investigation into the benefits and challenges associated with AI implementation in educational settings [9]. Furthermore, Ref. [10] recommends investigating the adoption of AI tools in education, emphasizing the importance of understanding the reasons behind the slower adoption in certain subjects to inform more effective implementation strategies.

The AI Index Report 2023 [11] shows that IPSOS conducted a wide-ranging public opinion survey on attitudes towards AI. The IPSOS survey makes cross-country comparisons to gauge respondents’ views on the potential benefits versus drawbacks of AI products and services. This offers valuable insights into the divergent attitudes towards AI in different global contexts, such as Saudi Arabia (SA), China, India, and the United States. The AI Index Report 2023 found that Saudi Arabia and India have very positive public opinions towards AI. A total of 76% of Saudi respondents and 71% of Indian respondents believed AI products have more benefits than drawbacks, the second and third highest percentages globally after China. In contrast, only 35% of Americans shared this favorable view. This suggests a pronounced difference in how the public in Saudi Arabia, India, and other countries, such as the United States, perceive AI technology, with Saudi Arabia and India exhibiting a much more optimistic stance. Overall, the findings from the IPSOS survey, as reported in the AI Index Report 2023, highlight the varying perceptions of AI across the globe and the need to consider cultural and national contexts when understanding public attitudes towards this emerging technology. The study of [12] provides valuable insights into the global adoption and usage of GenAI tools among students. According to the report, students in Saudi Arabia are among the global leaders in GenAI tool adoption, with 32% of students reporting weekly or more frequent usage. Moreover, a global student survey in 2023 by [13] found that 62% of undergraduate students in Saudi Arabia have used GenAI for their studies. These reports indicate a strong future outlook for GenAI tool usage among Saudi Arabian students. Furthermore, Saudi Arabia is actively exploring the potential of GenAI across various fields, including education, with the aim of harnessing its capabilities effectively [14]. In response to the growth of ChatGPT, the Saudi Data and AI Authority (SDAIA), the competent authority concerned with data and AI, has issued guidelines on “Generative Artificial Intelligence in Education” and developed a comprehensive “Saudi Academic Framework for AI Qualifications” to address the opportunities and challenges presented by this technology [14,15].

Remarkably, the existing literature shows significant differences in how students adopt GenAI tools across different countries [11,12,13,16]. More research is needed to understand what causes these differences, such as educational practices, technology, cultural attitudes, and institutional policies. As GenAI technology continues to develop, understanding how students engage with these tools worldwide will be essential for effectively integrating them into teaching, learning, and research. Ref. [17] suggests that using ChatGPT in higher education is a new and promising field that can drive innovation and support sustainable development in education, society, and the economy.

Therefore, the objective of this research is to investigate the current perceptions related to GenAI tools among higher education students in SA. Specifically, this study aimed to address the main research question: what are the current perceptions of higher education students towards GenAI tools? This investigation was guided by the following sub-research questions:

RQ1: To what extent do higher education students utilize GenAI tools?

RQ2: What benefits and challenges do higher education students encounter while using GenAI tools in education?

2. Background

2.1. GenAI Tools in Higher Education

The advanced capabilities of GenAI to handle complex prompts and produce human-like output have led to growing research and interest in integrating these technologies across a wide range of fields, including healthcare, medicine, media, and education [5]. This advancement has led to new opportunities for studying how students use and benefit from AI technologies in educational settings. Investigating the impact of GenAI on students’ learning experiences is a crucial area of scientific research in the field of AI in higher education [4,5].

Recent studies have extensively examined the role and impact of GenAI technologies in higher education. In [4], the researchers conducted a systematic review from 2016 to 2022, revealing a significant increase in publications during 2021 and 2022. This review identified key trends such as a shift in research focus from the US to China, greater involvement of researchers from education departments, and a primary focus on undergraduate students. It was found that 72% of the studies focused on students, 17% on instructors, and 11% on managers. Additionally, five key themes for AI in higher education were highlighted: assessment/evaluation, prediction, AI assistant, Intelligent Tutoring Systems (ITS), and managing student learning. The review also pointed out gaps in the literature and recommended further exploration of new tools like Chat GPT.

Focusing on GenAI technologies, Ref. [5] surveyed 399 undergraduate and postgraduate students from various disciplines in Hong Kong to understand their perceptions of ChatGPT in higher education, finding a generally positive attitude towards integrating GenAI in teaching and learning. Similarly, Ref. [18] surveyed 200 Vietnamese university students, revealing a generally positive attitude towards the integration of ChatGPT into their learning experiences. Additionally, Ref. [19] examined subgroups and potential systematic relationships between students’ backgrounds (e.g., gender, level of study, discipline, and university affiliations) and their response tendencies, providing a foundation for more informed discussions about AI’s role in higher education.

A large-scale quantitative study was conducted by [20], with 1135 participants from an Australian university, revealing that most students had limited knowledge, experience, and confidence in using GenAI tools. The research highlights the need for appropriate support and guidance in integrating GenAI tools like ChatGPT. Similarly, a nationwide survey was conducted by [21] involving over 6300 students across Germany to analyze the use and characteristics of AI-based tools, such as ChatGPT and GPT-4, among university students. These studies underscore the necessity for appropriate support and guidance in integrating GenAI tools into educational settings.

More recently, Ref. [22] focused on student perspectives at the University of Liverpool, surveying 2555 students to inform changes to the university’s Academic Integrity code of practice. Their findings showed that, while only 7% of students were unfamiliar with GenAI technologies, over half had used or considered using them for academic purposes. Most students supported tools like Grammarly but were opposed to using ChatGPT for writing entire essays, highlighting the need for clear policies on GenAI use to ensure equal access for all students.

GenAI offers various possibilities for its utilization in both undergraduate and graduate/postdoctoral research and writing. These include conducting literature reviews, summarising text, analyzing data, prototyping and simulations, communicating publicly, drafting and editing, coding, and generating thesis topics and ideas [14,23].

Table 1. Examples of used GenAI tools in education.

	Tools	Usage/URL
1	ChatGPT	The platform enables users to engage in conversations, gain insights, and automate tasks (https://chat.openai.com/, accessed on 1 August 2024).
2	CoPilot	The chatbot uses a large language model to cite sources, create poems, and write songs (https://copilot.microsoft.com/, accessed on 1 August 2024).
3	QuillBot	QuillBot provides AI-powered tools like a Grammar Checker, Paraphraser, AI Writer, and Summarizer (https://quillbot.com/, accessed on 1 August 2024).
4	Gemini	Gemini offers direct access to Google AI, assisting with writing, planning, learning, and research (https://gemini.google.com/, accessed on 1 August 2024).
5	ChatSonic	ChatSonic is an AI chatbot that uses advanced natural language processing technology to answer questions accurately and informatively (https://chatsonic.pro/, accessed on 1 August 2024).
6	Socratic	Socratic supports subjects like Science, Math, Literature, and Social Studies by providing visual explanations of key concepts (https://socratic.org/, accessed on 1 August 2024).
7	Tome	Tome is an AI tool that creates polished and professional presentations (https://tome.app, accessed on 1 August 2024).
8	Galileo AI	Galileo AI is a UI generation platform that facilitates easy and fast design ideation (https://www.usegalileo.ai/explore, accessed on 1 August 2024).
9	Midjourney	Midjourney creates high-quality, unique images using advanced AI (https://midjourney.com/, accessed on 1 August 2024).
10	GPTZero	GPTZero is an AI detection software designed to identify text generated by large language models (https://gptzero.me/, accessed on 1 August 2024).

presents some examples of GenAI tools that are commonly recognized within the higher education context [2,3,6,14,16,19,24,25].

2.2. Benefits and Challenges of GenAI in Higher Education

The integration of GenAI into education has attracted considerable attention, with various researchers exploring its potential benefits and concerns in higher education. GenAI offers several advantages for students in the teaching and learning process. It has enabled personalized and immediate learning support [5,17,26,27], revolutionizing the way students learn [5,17,26,28]. GenAI improves learning experiences by providing automated and personalized assistance and feedback [5,18,26,27], saving time [5,18], enabling access to information across different subject areas [18,29], offering quality explanations and clarifications [18,21,26], and promoting critical thinking skills [26]. Additionally, GenAI tools are considered useful research aids for brainstorming and generating writing ideas, analyzing data, creating visual and audio multimedia, and handling administrative tasks [5,18]. GenAI applications in programming assistance also have the potential to lower learning barriers, boost work efficiency, and minimize repetitive coding tasks [30]. Another significant benefit of GenAI is its application in teaching, where it can serve as an assessment tool and generate original content to enhance teaching methods [4,28].

On the other hand, the growing presence of AI technologies in classrooms presents researchers and educators with critical questions and challenges. Of particular concern are the limitations of GenAI and issues related to misinformation and hallucination, ethics, privacy, plagiarism, copyright, academic integrity, and preserving the authenticity of student work [23,26,28,29,31]. Additionally, there are challenges regarding the effective implementation of AI and its potential long-term impact on teaching roles [18]. Students have encountered various obstacles and weaknesses when using ChatGPT in the teaching and learning process, including limitations in response length, potential bias in results, inaccuracies in information, the possibility of forged citations and references, false statements and quotes from non-existent experts, challenges in evaluating source quality and reliability, accurately citing sources, effectively paraphrasing, and using idioms, as well as a lack of improvement in motivation for challenging tasks [18,23,26]. These challenges emphasize the importance of educating AI users about these risks and highlight the necessity of ethical practices. GenAI is best used to augment traditional teaching methods, enhancing interactivity and personalizing the learning experience [28].

2.3. Theoretical Framework

The theoretical foundations for this study draw from several key concepts and models related to technology adoption and educational uses of artificial intelligence (AI) tools. The Technology Acceptance Model (TAM) provides a robust framework for understanding students’ awareness, familiarity, and willingness to use GenAI tools in their academic work. TAM posits that perceived usefulness and perceived ease of use are primary determinants of an individual’s intention to use a new technology [32,33]. In the context of GenAI tools, perceived usefulness reflects how students believe these tools enhance their academic performance, while perceived ease of use relates to their comfort and familiarity in utilizing such technologies. By integrating TAM, this study aims to assess how these factors influence students’ adoption and frequent use of GenAI tools in their educational endeavors. Additionally, the Task-Technology Fit (TTF) theory is utilized to gain deeper insights into higher education students’ perceptions of GenAI tools and the benefits and challenges they experience in their learning. TTF suggests that the effectiveness of technology relies on the alignment between Task Requirements and Technology Functionality [34,35]. Task Requirements guide how students engage with their academic tasks, influencing their use of GenAI tools, while Technology Functionality pertains to the capabilities of these tools in supporting those tasks. The integration of these theoretical perspectives provides a comprehensive foundation for exploring the multifaceted aspects of GenAI tool usage among higher education students.

3. Methodology

3.1. Research Design

This study adopts a rigorous quantitative research methodology to provide a comprehensive, data-driven understanding of the current state of GenAI tool usage among higher education students. By utilizing a cross-sectional survey design, this research effectively examines participants’ attitudes, beliefs, opinions, and practices at a specific point in time [36]. This study takes a snapshot of the participants’ responses during the second semester of the 2023/2024 academic year, specifically from 17 March 2024 to 17 June 2024. The descriptive methods include surveying a representative sample of higher education students to measure their utilization of GenAI tools, analyzing the survey data to describe the educational purposes for which students use GenAI tools, and identifying the perceived benefits and challenges associated with the use of GenAI tools in education. By employing both descriptive and analytical research methods, this study aims to provide a comprehensive, data-driven understanding of the current state of GenAI tool usage among higher education students, which can inform the development of relevant policies, curricula, and support systems.

3.2. Research Context

The current study was conducted at the main campus of King Saud University (KSU) in Riyadh, Saudi Arabia. As one of the oldest and largest universities in the country, KSU is deeply engaged in AI research and education [37]. The university provides a range of programs and courses in computer science, data science, and various AI-related disciplines [38].

All students at KSU undergo the Common First Year (CFY) program, which serves as a foundational academic curriculum aiming to provide fundamental knowledge across diverse disciplines. This program acts as a transition phase between secondary education and specialized undergraduate studies, offering core courses in mathematics, natural sciences, humanities, social sciences, and English language skills. Participation in the CFY program allows students to explore different academic fields, identify their interests, and prepare for their chosen majors [39]. Following the CFY program, students can select one of KSU’s main colleges, each dedicated to various fields of study [39], including the following:

• Health Colleges include six disciplines: Medicine, Dentistry, Pharmacy, Applied Medical Sciences, Nursing, and Prince Sultan Bin Abdulaziz College for Emergency Medical Services.
• Science Colleges include six disciplines: Computer and Information Sciences, Business Administration, Architecture and Planning, Food and Agricultural Sciences, Sciences, and Engineering.
• Humanities Colleges encompass seven disciplines: Education, Law and Political Sciences, Language Sciences, Tourism and Archaeology, Sport Sciences and Physical Activity, Arts, and Humanities and Social Sciences.
• Community Colleges consist of two disciplines: Community College and College of Applied Studies and Community Services.

3.3. Research Sample

The population for this study consisted of higher education students enrolled at KSU in SA. According to the 2023 statistics from KSU, the university has a total enrollment of 54,292 students, of which 30,581 are male and 23,711 are female [40]. Utilizing a comprehensive survey technique to ensure broad participation, the survey was distributed to the entire KSU community, resulting in a total of 859 responses. This sampling approach ensured that the survey respondents were representative of the diverse student body at KSU, capturing perspectives from both male and female students across different academic programs and year levels. The demographic information of the participants is presented in

Table 2. Demographic information of participants (n = 859).

Attributes	Frequency	Percentages
Gender
male	501	58.3%
female	358	41.7%
Academic level
diploma	34	4%
undergrad	598	69.6%
master	123	14.3%
PhD	104	12.1%
College
Common First Year (CFY)	114	13.3%
Health Colleges	116	13.5%
Science Colleges	346	40.3%
Humanities Colleges	253	29.4%
Community Colleges	30	3.5%
Total	859	100%

.

These demographic characteristics of the sample provide important contextual information for interpreting the research findings. The overrepresentation of male participants, undergraduate students, and those from scientific colleges should be considered when generalizing the results.

3.4. Research Instrument

A survey questionnaire (Supplementary Materials) was used to collect data, making it suitable for large sample sizes and measurable variables, allowing efficient data collection from a broad population while maintaining ethical distance [41]. This study’s instrument, developed through literature review and expert consultations, consisted of a customized questionnaire based on existing research on AI tools in education [6,16,18,19,42,43,44]. It included two sections: demographic information and students’ perceptions of GenAI tools, with 38 items covering utilization, purposes, benefits, and challenges. The questionnaire contained an initial question on GenAI utilization, classifying respondents as users or non-users, with non-users asked for reasons and users providing details about usage, benefits, and challenges. The questionnaire was in Arabic, with clear instructions, including an estimated completion time.

The survey design effectively integrates both the Technology Acceptance Model (TAM) and the Task-Technology Fit (TTF) framework. For instance, TAM is represented in the Awareness and Familiarity section by the question, “Do you use generative AI tools in higher education?”. Additionally, the Educational Purposes question “I use GenAI tools in education (choose all that apply)” aligns with TAM’s constructs of Perceived Ease of Use and Perceived Usefulness, as well as TTF’s focus on Task Requirements. Furthermore, TTF is illustrated through perceived benefits, such as “enhance academic performance”, and challenges encountered when using GenAI tools, like “provide inaccurate or false references”, which correspond to Technology Functionality.

To determine validity, the questionnaire was reviewed by four faculty members, leading to revisions based on their feedback. The internal consistency was high, with a Cronbach’s alpha of 0.79. Additionally, a mini-pilot was conducted with a small group of students before the full survey rollout to identify and refine any confusing phrasing, ensuring greater clarity and accessibility for the student population. The review and pilot study refined the questionnaire to 28 items by revising or eliminating 10 items that were redundant, unclear, or misaligned with the study’s objectives. Also, the response format for the question on “the educational purposes of employing GenAI” was updated from a Likert scale to a “choose all that apply” format, which reduced the number of questions in the questionnaire.

The survey was administered electronically via Google Forms to KSU students during the second semester of the 2023/2024 academic year. It was initially sent via email with a study explanation and survey link. To boost responses, SMS reminders, on-campus paper slips, and social media posts were also utilized. All statements were made mandatory, and responses from non-KSU students were excluded. Therefore, out of the 869 responses received, 10 were excluded as they were from non-KSU students.

3.5. Data Analysis

Quantitative data from the survey were analyzed using descriptive statistics, including percentages, means, and standard deviations, to summarize student utilization, benefits, and challenges associated with GenAI tools in higher education. The SPSS (V25) software was employed to compute frequencies and percentages for each survey item. Data were initially exported from Google Forms in Excel format, then imported and reformatted in SPSS for detailed analysis.

3.6. Ethical Considerations

The Ethics of Human and Social Research Committee at KSU approved the ethical conduct of this study (KSU-HE-24-296). All participating students provided informed consent before joining the study. They were also made aware of their anonymity and their right to withdraw from the study at any time without needing to provide a reason

4. Findings and Discussions

4.1. RQ1: To What Extent Do Higher Education Students Utilize GenAI Tools?

The present study aimed to investigate the extent to which higher education students utilize GenAI tools. This study revealed that, among the 859 higher education students surveyed, the majority used GenAI tools, while a notable proportion did not use these tools (see Figure 1). The following sections detail the reasons why some students do not use GenAI tools and explore the applications and purposes of using GenAI among students who do use these tools.

4.1.1. Students Who Do Not Use GenAI

The findings from this study provide valuable insights into the reasons why a sizable minority (21.3%, n = 183) of higher education students do not utilize GenAI tools. The survey offers five choices for reasons not to use GenAI tools in higher education: “I have limited knowledge of GenAI tools”, “To avoid the harms of using GenAI tools”, “I lack interest in GenAI tools”, “I have not heard about GenAI tools”, and an option for respondents to provide other reasons. The most commonly cited reason was having limited knowledge about these technologies (n = 80). This suggests that a significant proportion of the student population may be unaware of the potential applications and benefits of these emerging technologies within the educational context. Additionally, some respondents (n = 75) indicated that they had not heard about GenAI tools, further underscoring the need for greater awareness and education about these technologies among the student community. These data also reveal that a smaller, but still substantial, proportion of students (n = 38) expressed a lack of interest in using GenAI tools, while (n = 29) reported concerns about the potential harms associated with these technologies.

Consistent with these findings, Ref. [5] results reported that 33.3% of participants had never used GenAI technologies, suggesting some hesitancy or resistance to engaging with these emerging tools. This aligns with the research of [45], which suggested that some students may be less inclined to use new technologies due to a lack of digital literacy or technological proficiency. Similarly, a study examining university students’ perceptions of ChatGPT, with the goal of informing updates to the University of Liverpool’s Academic Integrity code of practice, found that only 7% of responding students were unaware of any GAI technologies [22]. Additionally, Ref. [20] surveyed 1135 students from an Australian university and found that most students had little knowledge, experience, and confidence in using GenAI tools. Furthermore, 13% of students had either heard nothing or very little (28%) about ChatGPT and other GenAI tools. These findings, taken together, suggest that while GenAI technologies are becoming more prevalent, a significant portion of the student population may still lack familiarity, experience, and confidence in using these emerging tools.

The open-ended part of the reasons for not using GenAI tools provided further perceptions from a sample of 16 students. For example, the reasons of two students were, “In the engineering specialty, there are unlimited ideas that are difficult for artificial intelligence to help me with, and it gives me flat, useless answers”, and “AI tools are still not advanced in the humanities and may provide false information”. These perceptions highlight the students’ concerns about the limitations of GenAI in engineering and the humanities. They may be wary of relying on AI-generated information, as they may perceive it as potentially inaccurate or misleading, especially in fields where subjective interpretation plays a significant role.

Other students stated that “People are smarter than these tools, and I would not like them to work for me, I can do what they do better” and “I feel that I can always find what I need on my own without resorting to artificial intelligence tools”. These perceptions reflect a strong sense of self-reliance and confidence in their own intellectual abilities compared to GenAI tools. Students feel they can produce better work on their own and achieve the same or better results through their own research and problem-solving without needing to rely on GenAI.

These perceptions suggest resistance to incorporating GenAI into academic workflows, as students place a high value on their own intelligence and problem-solving skills. This belief in their superiority over AI may stem from a lack of familiarity with the evolving capabilities of these tools, as well as a preference for traditional research methods. Addressing these underlying perceptions will be key to helping students recognize the potential benefits of integrating GenAI to complement their strengths rather than viewing it as a replacement for human intelligence.

4.1.2. Students Who Are Using GenAI

This study revealed that the majority of higher education students, comprising 78.7% (n = 676), frequently used GenAI tools, aligning with the findings of [22], which found that 68.9% of students were familiar with ChatGPT. The findings also revealed that 18.2% reported rare usage, 38.5% used these tools sometimes, and 22.0% used them often. This variation highlights the diverse levels of engagement with GenAI tools among students. Furthermore, the findings provide valuable insights into the integration of GenAI tools within the higher education context.

Table 3. Using GenAI tools in the educational environment.

	Frequency	Percentages
Widespread among my peers	528	78.1%
My teachers (s) encourage its use	178	26.3%
My university has established rules or guidelines for its responsible use	130	19.2%

data demonstrate widespread adoption among students, with 78.1% indicating common usage among their peers. This widespread adoption among students highlights the growing prominence and perceived utility of these emerging technologies in academic settings. Additionally, the data reveal that a sizable proportion of students (26.3%) report that their instructors actively encourage the use of GenAI tools. This suggests that a significant number of educators recognize the potential benefits and applications of GenAI in enhancing learning, collaboration, and academic productivity. Similar to these findings, Ref. [46] emphasizes that higher education teachers should be knowledgeable about the capabilities of ChatGPT.

The data also indicate that only 19.2% of students are aware of their university’s established rules or guidelines for the responsible use of these technologies. This relatively low awareness suggests a need for better communication and dissemination of this information among students. Enhancing awareness and understanding of institutional guidelines could foster a culture of responsible GenAI use. This would empower students to leverage these technologies effectively while maintaining the university’s academic standards and ethical principles. Other studies have also highlighted low awareness of rules or guidelines regarding the responsible use of AI. For instance, Ref. [19] found that 55% of the 5894 higher education students surveyed across Swedish universities were unsure about the rules or guidelines for responsibly using chatbots provided by their teachers and university.

Moreover, the findings provide a comprehensive overview of the GenAI tools utilized by higher education students. The data presented in

Table 4. GenAI tools used by higher education students (n = 676).

GenAI Tools	Frequency	Percentages
ChatGPT	583	86.2%
Gemini	148	21.9%
Socratic	114	16.9%
CoPilot	111	16.4%
Quill Bot	104	15.4%
Midjourney	39	5.8%
GPTZero	36	5.3%
Tome	31	4.6%
Chat Sonic	27	4.0%
Galileo AI	15	2.2%

reveal that ChatGPT is the most widely adopted GenAI tool, with 86.2% of respondents reporting its use. This finding is similar to other studies that found ChatGPT to be the most used AI-based tool among higher education students [21,22,30]. Beyond ChatGPT, the data indicate that a sizable proportion of students have also explored other GenAI tools, such as Gemini (21.9%), Socratic (16.9%), and CoPilot (16.4%). The findings indicate that students are actively exploring a variety of Generative AI tools, likely seeking to leverage the distinctive capabilities and functionalities provided by these emerging technologies to enhance their academic and creative work.

It is noteworthy that a small but significant percentage of students (5.8%) have utilized Midjourney, a GenAI tool specialized in generating images, and GPTZero (5.3%), a tool designed to detect AI-generated text. This suggests that students are not only consuming Generative AI outputs but are also exploring ways to evaluate and understand the provenance and authenticity of such content, reflecting a growing awareness of the potential implications and ethical considerations surrounding the use of these technologies. Furthermore, the results from open-ended questions (n = 41) provide another perspective on GenAI tools used by higher education students, with (2.7%) using AI Poe, while some used various tools like Canva AI, Notion AI, and ChatPDF, representing (3.4%) of respondents.

In addition, the findings provide valuable insights into the diverse ways higher education students use GenAI technologies to support their academic tasks. The data presented in

Table 5. Educational purposes of employing GenAI among higher education students.

Educational Purposes of Employing GenAI	Frequency	Percentages
to define or clarify concepts.	468	69.2%
for translation	343	50.7%
to search for academic literature	282	41.7%
to summarize articles, books, and videos.	310	45.9%
to generate ideas while writing	360	53.3%
to enhance the quality of writing	276	40.8%
to proofread and edit my writing	278	41.1%
to assist in completing assignments.	280	41.4%
to assist in completing home exams.	115	17.0%
to facilitate project work.	321	47.5%
to assist in creating digital multimedia and presentations.	142	21.0%
to aid in solving numerical problems.	125	18.5%
to support coding	146	21.6%

indicate that the most common use of these tools is for defining or clarifying concepts, with 69.2% of students reporting this application. Additionally, 50.7% of students use GenAI for translation purposes. This interesting finding shows that GenAI tools allow students to efficiently process and comprehend course materials and research sources in different languages. This suggests that students utilize GenAI to enhance their understanding of course material and build a stronger foundation of knowledge. Previous studies have found that GenAI helps international students and non-native English speakers clarify meanings and understand complex concepts through transcription and translation [16,24,47].

Additionally, a significant proportion of students employ GenAI to generate ideas while writing (53.3%) and to summarize academic literature (45.9%), highlighting the potential of these technologies to streamline research and writing processes. The data also show that students are using GenAI to search for relevant sources (41.7%), enhance the quality of their writing (40.8%), and improve proofreading (41.1%). These findings align with [5,10,47], who found that higher education students use GenAI technologies like ChatGPT, Grammarly, and QuillBot for learning, writing, and research purposes, including ideas generation, literature searching, summarizing readings, grammar checking, brainstorming, paraphrasing, and generating hypotheses from data analysis.

Furthermore, the data reveal that students are utilizing GenAI to assist in completing assignments (41.4%) and home exams (17.0%), as well as to facilitate project work (47.5%) and the creation of digital multimedia and presentations (21.0%). This suggests that these emerging technologies are being integrated into a wide range of academic tasks and activities, potentially improving efficiency, productivity, and the overall quality of student work. Similar findings were reported by other studies [5,16,19,47], highlighting GenAI’s ability to deliver personalized and adaptive learning experiences tailored to students’ needs, preferences, and learning styles [5,19,24,43,44].

Interestingly, a smaller proportion of students (18.5%) report using GenAI to aid in solving numerical problems, and 21.6% use these tools to support coding, indicating that the application of GenAI in certain computing-related domains may be less prevalent or still in the early stages of adoption. Similarly, Ref. [30] found that the acceptance of AI Coding Assistant Tools among Chinese students is currently limited, and the factors affecting this acceptance are not well understood. Additionally, students at Manchester University use GenAI, particularly ChatGPT, to solve maths problems and develop algorithms, but they are often confused by multiple answers and concerned about the reliability of coding responses [16]. The capabilities of GenAI, such as proficiency in producing code, solving mathematical equations, and designing scientific experiments, may appear to be more directly practical and applicable to students specializing in science and engineering [20].

4.2. RQ2: What Benefits and Challenges Do Higher Education Students Encounter While Using GenAI Tools in Education?

The findings show the perceived benefits that higher education students associate with the use of GenAI tools.

Table 6. Perceived benefits of using GenAI tools in education among higher education students.

Perceived Benefits		Disagree	Neutral	Agree	Mean	Std. Deviation
1. easy to access and use.	F	9	54	613	2.89	0.35
	%	1.3	8.0	90.7
2. save time on tasks	F	8	46	622	2.91	0.33
	%	1.2	6.8	92.0
3. provide instant feedback	F	23	171	482	2.68	0.54
	%	3.4	25.3	71.3
4. foster critical thinking and problem-solving	F	98	206	372	2.41	0.73
	%	14.5	30.5	55.0
5. enhance general language ability	F	87	193	396	2.46	0.71
	%	12.9	28.6	58.6
6. enhance academic performance	F	57	183	436	2.56	0.64
	%	8.4	27.1	64.5
7. improve learning engagement	F	76	155	445	2.55	0.69
	%	11.2	22.9	65.8
8. increase confidence while learning	F	77	166	433	2.53	0.69
	%	11.4	24.6	64.1
Mean					2.62

presents eight statements capturing these benefits, including ease of access and use, time-saving, instant feedback, enhancement of critical thinking and problem-solving, language ability improvement, academic performance enhancement, learning engagement, and increased confidence. The standard deviation values suggest a moderate level of variability in the responses, indicating that, while the majority of students agree with the benefits, there is still some diversity in their perceptions. Over 55% of students agree with the positive aspects of GenAI tools, particularly in terms of ease of access and use (90.7% agree), time-saving (92.0% agree), and providing instant feedback (71.3% agree). Refs. [18,42] support this by highlighting that students view AI as a valuable tool for enhancing efficiency and saving time. Similarly, Ref. [4] found that GenAI tools support assessment and student learning management, while [47] reported that these tools are efficient and effective, with 50% of students responding positively to AI’s ability to provide customized instruction and feedback.

However, the agreement levels are lower for more complex benefits, such as fostering critical thinking and problem-solving (55.0% agree), enhancing general language ability (58.6% agree), and improving academic performance (64.5% agree). While these findings align with [47] regarding AI’s positive impact on higher education students’ linguistic capabilities, they contrast with [47] as well, where 25% of students using AI for learning expressed concerns that AI might negatively affect their learning outcomes. These students also believed that overreliance on AI tools could hinder the development of critical thinking and problem-solving abilities.

The overall mean agreement score across all the statements is 2.62, indicating a generally positive perception of the benefits of GenAI tools among higher education students. This aligns with [5,10,19], who reported positive perceptions of AI in teaching and learning among most students, with [10] also highlighting ChatGPT’s extensive capabilities in facilitating problem-solving and academic research.

Furthermore,

Table 7. Challenges of using GenAI tools in education among higher education students.

Challenges		Disagree	Neutral	Agree	Mean	Std. Deviation
1. require fast internet connection	F	222	206	247	2.04	0.83
	%	32.8	30.5	36.5
2. require a subscription fee for accessing advanced features	F	102	181	392	2.43	0.74
	%	15.1	26.8	58.0
3. poses a risk to privacy and data security	F	164	315	196	2.05	0.73
	%	24.3	46.6	29.0
4. provide unreliable information	F	88	307	280	2.28	0.68
	%	13.0	45.4	41.4
5. provide inaccurate or false references	F	94	295	286	2.28	0.69
	%	13.9	43.6	42.3
6. restrict learning autonomy and narrow learning experiences	F	147	296	232	2.13	0.74
	%	21.7	43.8	34.3
7. reduce human-to-human interaction	F	148	213	314	2.25	0.79
	%	21.9	31.5	46.4
8. lead to plagiarism and cheating	F	161	226	288	2.19	0.79
	%	23.8	33.4	42.6
9. negatively impact learning in the future	F	208	224	243	2.05	0.82
	%	30.8	33.1	35.9
Mean					2.19

shows the findings related to the challenges that higher education students encounter while using GenAI tools in education. The overall mean score of 2.19 suggests that students perceive these challenges as relatively significant, with a tendency towards agreement. This indicates that the issues highlighted are not merely isolated concerns but rather represent widespread sentiments among the student population. One of the most pressing challenges, as evidenced by the high mean score of 2.43, is the requirement of a subscription fee to access advanced GenAI features. This financial barrier can severely limit the accessibility of these tools for students, particularly those from low-income backgrounds or with limited financial resources. This finding underscores the importance of ensuring equitable access to educational technologies. The study of [43] notes that AI systems can be costly to implement and maintain, posing a challenge for budget-constrained universities. Therefore, universities must carefully consider the costs and benefits of integrating AI into their classrooms.

Another area of concern is the reliability and accuracy of the information provided by GenAI tools. The high mean scores for “provide unreliable information” (2.28) and “provide inaccurate or false references” (2.28) suggest that students are wary of the trustworthiness of the outputs generated by GenAI. This finding aligns with [24], which found that GenAI can lead to the inclusion of irrelevant or inaccurate information in students’ essays. Similarly, Ref. [16] reported several concerns among higher education students about using AI in education, including worries about information accuracy, privacy and data security, lack of regulations, teacher use of AI, fairness in access, and over-reliance on AI. This emphasizes the need for a careful and inclusive approach to integrating AI technologies into learning. Additionally, Ref. [47] reported that 75% of interviewed students encountered mistakes in AI outputs, noting that AI tools are not infallible and sometimes provide incorrect or misleading information

The table also highlights concerns about the impact of GenAI tools on learning autonomy and the breadth of learning experiences. The mean score of 2.13 for the challenge “restrict learning autonomy and narrow learning experiences” suggests that students are worried about the potential for these tools to limit their ability to engage in self-directed and diverse learning activities. Institutions should consider ways to ensure that the integration of GenAI tools into the curriculum preserves and enhances, rather than diminishes, the opportunities for autonomous and multifaceted learning. Additionally, the table reveals concerns about the potential for GenAI tools to reduce human-to-human interaction (mean = 2.25) and lead to plagiarism and cheating (mean = 2.19). Similarly, Ref. [24] noted that using ChatGPT increases the likelihood of plagiarism and presents challenges in assessing students’ work. In line with these findings, Ref. [5] reported that the key challenges of GenAI use include concerns about accuracy and transparency, privacy and ethical issues like plagiarism, and the potential to hinder individual development and holistic competencies through over-reliance on the technology. These challenges underscore the importance of maintaining the social and collaborative aspects of education and the need for robust academic integrity measures to prevent misuse of these technologies. In addition, these findings align with the literature that emphasizes the critical need for a thoughtful, deliberative, and cautious approach to the utilization of AI tools, where academic rigor, ethical considerations, and human intelligence are centrally positioned and prioritized throughout learning [46].

5. Conclusions

Overall, this study provides valuable insights into the ways in which higher education students perceive and utilize GenAI tools, which can inform the design and implementation of educational policies and practices related to the adoption and integration of these technologies in the academic context. This study revealed that 78.7% of higher education students frequently use GenAI tools, consistent with the finding of [22] that there was 68.9% familiarity with ChatGPT among students. However, 21.3% of students do not use GenAI tools, primarily due to a lack of knowledge or interest. Among those who do use these tools, the primary purposes include defining or clarifying concepts, translation, generating ideas in writing, and summarizing academic literature.

Grounded in the Technology Acceptance Model (TAM) and the theory of Task-Technology Fit (TTF), this research provides a comprehensive understanding of the complex factors shaping students’ views on the integration of AI technologies. The TAM framework suggests that students’ perceptions of the usefulness and ease of use of GenAI tools are key determinants of their intention to adopt and utilize these technologies. The concerns raised, such as worries about the accuracy of information, learning autonomy, plagiarism, and reduced human interaction, can be interpreted through the lens of perceived usefulness. Addressing these concerns may be crucial for enhancing students’ acceptance and adoption of GenAI tools in their academic activities. The TTF theory emphasizes aligning GenAI tool capabilities with students’ educational needs. The findings indicate that students see the potential to enhance understanding, learning, and research processing. However, concerns remain about a mismatch between tool capabilities and requirements. Additionally, the research suggests that GenAI applications may be less prevalent or in the early stages in certain computing-related domains. By addressing alignment between AI tools and diverse student needs across disciplines, universities can unlock the potential of these technologies to improve learning and research. By drawing on the complementary perspectives of TAM and TTF, this study provides a robust theoretical foundation for understanding the multifaceted aspects of GenAI tool usage among higher education students.

Despite this study’s valuable insights, it faced several limitations. It is important to note that the majority of our sample consisted of male students and those from science colleges. This composition may limit the generalizability of our findings and could potentially misrepresent the views of female students and those from non-science colleges. Additionally, the research context, such as institutional policies and technological infrastructure, may have influenced the findings, limiting transferability. This study provided a snapshot in time without exploring how student perceptions might shift as they gain more experience with the advanced technology, highlighting the need for a longitudinal perspective. Future research could adopt more longitudinal approaches to better understand how student perceptions and experiences with GenAI evolve over time and across diverse institutional contexts. Additionally, studies that closely monitor the rapid technological advancements in GenAI would help ensure the ongoing relevance and applicability of research findings.