Recently, in the 21st century, accomplishing learning goals is increasingly difficult. Even though generative AI technology is widely available, the reality is that the skills needed still need improvement. Moreover, generative AI literacy, including knowledge of generative AI applications and tools, must be included as an axis to prepare students to use generative AI. Hence, 21st century skills include communication, critical thinking, collaboration, and creativity. Students with these skills can lead change to keep pace with scientific and technological progress and confront life problems that students face while they learn. This study supports students in developing their future skills and will guide educators and policymakers to adapt their teaching strategies and policies in higher education institutions. Moreover, this study’s outcomes can benefit the advancement of generative AI in education for college instructors, researchers, and policymakers. Theoretically, this study is unique in its investigation since AI literacy and 21st century skills that might influence generative AI acceptance have been rarely explored by other researchers.

Therefore, the importance of the study stems from theoretical and applied considerations. It highlights the importance of AI literacy and 21st century skills in accepting generative AI applications in higher education. Also, this study contributes to closing the gap in AI literacy and 21st century skills that was merely investigated by applying the technology acceptance model (UTAUT) to determine how these factors influence students’ acceptance of using generative AI in learning. Additionally, the results and recommendations of this study might help university policy makers and college instructors to enhance the acceptance of generative AI.

Educational institutions, especially universities, keep pace with scientific and technological development by recognizing the importance of generative AI in education. In line with the recommendations of many researchers, it is necessary to explore AI literacy and 21st-century skills to improve and develop the educational process. It is necessary to enhance its competitive advantage to keep pace with such developments and benefit from them in teaching. This requires paying attention to students’ needs and preparing them to possess the necessary skills and experience to accept this technology. Equally important, more attention should be given to preparing a generation that can deal with the challenges of the times, be aware of the potential of this technology, and invest in generative AI. Living in the 21st century requires students to deal safely and effectively with generative AI data. Future 21st-century skills qualify and enable students to accept this modern technology.

AI will influence various aspects of human life beyond just the computer industry, making it essential for everyone to understand AI. Today, AI is used across diverse sectors such as business, science, art, and education to enhance user experience and boost efficiency (Ng et al., 2021). AI applications, including smartphones and virtual assistants, are integrated into many aspects of our daily lives. People recognize AI services and devices; however, they often lack knowledge about the underlying concepts and technologies, and awareness of potential usage of AI (Ng et al., 2021). As learning becomes more infused with AI tools, it is crucial to support AI literacy for all (Sperling et al., 2024). AI literacy has become an essential skill set that everyone should acquire in response to this new era of technology advancement.

Moreover, 21st-century skills are needed to solve practical problems using generative AI. AI. Hence, this study is partially similar to Jing et al. (2024), Schiavo et al. (2024), and Strzelecki (2024), who explore AI literacy and generative AI adoption. Based on the foregoing, this study aims to explore college students’ AI and their 21st century skills influence on generative AI acceptance.

The ability to critically assess generative AI tools and apply them in a safe, ethical, and successful manner in various settings, including personal, professional, and educational ones, is known as AI literacy. Various measures of AI literacy have been employed to comprehend these elements. Some AI tools have been designed to assess these competencies in a particular population, such as college students (Tseng et al., 2025). Measures for computer science novices were created in other studies (Laupichler et al., 2023). Offered a wider variety of assessments considering AI literacy elements’ psychological and social components. As far as the researchers know, few studies have investigated the influence of AI literacy and 21^st-century skills on generative AI acceptance. At the same time, conducting this study is necessary as it is crucial to assess these variables in a developing country like Palestine. Also, this study addresses gaps tied to the study’s questions by limited focus on developing countries. Most studies (e.g., Kong et al., 2024) focus on high-resource settings, neglecting regions like Palestine. Also, demographic nuances, few studies explore how gender, field of study, and academic level intersect with AI acceptance (e.g., Stöhr et al., 2024). Moreover, existing research (e.g., Zhu et al., 2025) lacks longitudinal insights into how AI literacy evolves with using generative AI.

Limited infrastructure and resources may hinder AI integration (e.g., lack of standardized ICT policies, as noted by Ndibalema, 2025). Gender disparities in AI acceptance were observed, aligning with Jang et al. (2022). Compared to Malaysia, Mansoor et al. (2024) found higher AI literacy in Malaysian students due to robust STEM policies, while in Sub-Saharan Africa, Ndibalema (2025) highlighted challenges similar to Palestine (e.g., digital gaps). In Europe, Strzelecki and ElArabawy (2024) noted higher acceptance in Poland due to better institutional support.

Thus, this study attempts to identify the influence of AI literacy and 21st-century skills by addressing the following questions:

What is the overall state of AI literacy among Palestine Technical University-Kadoorie (PTUK) students?

AI literacy is acknowledged as an essential skill in the modern era of AI. Traditionally, literacy encompasses specific ways of thinking about and participating in reading and writing to comprehend or communicate ideas within a particular context (Laupichler et al., 2023). It involves maintaining a balanced perspective on technology to ensure responsible and healthy usage, understanding and addressing issues related to privacy, security, legal and ethical concerns, and the societal role of digital technologies (Khoo et al., 2024). Literacy is closely tied to knowledge, while competency refers to the ability to perform tasks successfully. Literacy centers on understanding, while competency highlights the practical application of knowledge with confidence. Furthermore, AI competence is an essential skill that enables users to interact with AI-driven applications while maintaining a reliable understanding of the underlying algorithms (Ng et al., 2021). AI literacy, on the other hand, can be defined as the ability identify, use, and critically assess AI products while adhering to ethical principles (Wang et al., 2009).

Ng et al. (2023) developed an AI Literacy scale that outlines four key constructs: using and applying AI tools, understanding what AI is, recognizing when AI is integrated into a system, and AI ethical awareness. Even though some studies investigated AI literacy, most conducted qualitative research methods focusing on exploring initial investigations (Ng et al., 2023). These studies have focused on improving AI literacy rather than quantifying it with other variables. Long and Magerko (2020) describe AI literacy as a variety of competencies that motivate students to evaluate AI applications critically, communicate and cooperate effectively with AI tools, and use AI as a system across diverse contexts. They indicate that AI literacy has 17 skills, highlighting its connections to digital data and computational literacy. While these literacies may be interdependent, they remain distinct. For example, AI literacy builds on basic computer skills, making digital literacy a foundational requirement. Data literacy, which involves the ability to understand, analyze, evaluate, and debate data, significantly overlaps with AI literacy due to the integral role of data in machine learning.

Later, Ng et al. (2023) expanded on this concept by framing AI literacy around four core dimensions. The first dimension, comprehension, involves grasping fundamental AI concepts, such as how AI operates, machine learning algorithms, data training, and AI biases. The second dimension focuses on using AI tools to solve problems and accomplish objectives, which often requires coding and the capacity to handle large datasets. The third dimension is evaluation, which entails assessing AI applications’ quality and reliability and ethically designing and developing AI systems. This demands technical skills and an awareness of AI’s ethical and societal impacts. The final dimension is AI ethics, which centers on understanding the moral and ethical considerations surrounding AI, enabling individuals to make informed decisions about its use. This includes issues such as fairness, transparency, privacy, and the broader societal implications. For example, Chan and Lo (2025) highlighted the proliferation of AI technologies such as facial recognition and predictive policing has exposed significant gaps in legal protections and ethical frameworks, particularly concerning privacy and autonomy. Also, their study warns against the erosion of privacy norms through continuous monitoring. To mitigate these risks, higher education institutions must adopt privacy-by-design principles and ensure algorithmic transparency in AI tools.

Moreover, Tseng et al. (2025) assessed AI literacy level of nursing students by using AI literacy scale that includes four categories: using and applying AI ability, understanding AI, detecting AI, making ethical considerations of AI by analyzing the influence of ChatGPT by Openia and Copilot tools compared with traditional teaching methods and found that generative AI tools enhanced students’ AI literacy.

Meanwhile, a study conceptualized AI literacy as a person’s capability to understand how AI technology functions and influences society comprehensively. It also involves using these technologies ethically. Similarly, Salhab (2024) explored AI literacy in the college curriculum and found that integrating AI literacy into the curriculum can enhance AI literacy for college students to foster essential 21st century skills. From a broader perspective, Mansoor et al. (2024) investigated AI literacy levels by surveying university students across four countries in Asia and Africa. They found variations in AI literacy levels influenced by nationality, field of study, and academic specialization, while gender did not significantly influence AI literacy. Malaysian participants demonstrated higher AI literacy levels compared to other students. Results also revealed that demographic and academic variables significantly shaped participants’ perceptions of AI literacy.

The 21st century skills equip students to develop different thinking types, make sense in learning experiences, and employ approaches that can be applied in many different life situations. 21st-century skills encompass collaboration, communication, critical thinking, and creativity (Kain et al., 2024). These learning skills were developed by educators, government officials, and business leaders to equip students for a constantly changing life and work environment (Cristea et al., 2024). Collaboration involves the ability to cooperate effectively and respectfully with diverse teams and flexibility and willingness to achieve a common goal. It assumes shared responsibility, collaborative work, and valuing each team member’s contributions (Barrett et al., 2021; He and Chiang, 2024). Communication includes the ability to listen and interpret meaning effectively. Strong communication skills are crucial for sharing knowledge, principles, feelings, and objectives. They are essential for informing, instructing, motivating, and persuading. Creativity is characterized by creating ideas, such as brainstorming, to generate valuable ideas. Students elaborate, refine, analyze, and evaluate original ideas to enhance and maximize creative thinking (Saleem et al., 2024). Critical thinking includes the ability to reason effectively, utilize systems thinking, make decisions, and solve problems (Liu et al., 2021).

These skills are important for reaching diverse audiences and conveying information to multilingual and multicultural populations (Shadiev and Wang, 2022). Moreover, a wide range of knowledge, capabilities, and work habits, like creative thinking, problem solving, innovation, and creativity skills, were practically operationalized in a specific setting, like a project-based STEM classroom using AI. For example, Hu (2024) documented a 6-month blended learning program where students used collaborative platforms (e.g., Padlet, GitHub) to solve community-based problems, illustrating applied critical thinking and creativity.

Over the past decade, educational systems have focused on helping students acquire these 21st century skills through various strategies and new technologies. These strategies include inquiry-based learning, AI-powered simulations, and gamification (Celik et al., 2024; Jing et al., 2024; Samala et al., 2024). Research literature shows that many studies have investigated these skills in light of using AI like robotics (Gratani and Giannandrea, 2022), and specific approaches like game-based learning and stimulating learning environments to develop these skills.

The extensive adoption of technology has led to an increasing preference for generative AI applications in education (Choung et al., 2023). Generative AI is an emerging technology comprising machine learning algorithms that produce original content like text, images, and sound (Alier et al., 2024). However, the implementation success of generative AI in educational settings largely hinges on students’ readiness to accept and embrace the technology (Li et al., 2024; Yilmaz et al., 2023). The UTAUT model provides insights into attitudes and intentions regarding using generative AI applications in education. For instance, students are more likely to use generative AI applications if they perceive them as intuitive and easy to use. Strzelecki and ElArabawy (2024) used UTAUT to investigate the key factors influencing students’ usage of ChatGPT. The authors reported that performance expectancy, effort expectancy, and social influence significantly impact behavioral intention to adopt ChatGPT. The study explores two factors that might affect students’ acceptance of generative AI, specifically within higher educational settings. Moreover, previous studies have shown that the UTAUT is an effective framework to evaluate users’ acceptance of modern technologies (Teng et al., 2022; Ustun et al., 2024). Therefore, the UTAUT could offer insights into how users evaluate these technologies. There is a noticeable gap in research grounded in theoretical frameworks like the UTAUT model that examines students’ acceptance of generative AI.

Generative AI applications can transform the roles and responsibilities assigned to students. The widespread adoption of these technologies can offer students enhanced access to information, deeper understanding of their learning progress, and more tailored educational experiences (Yilmaz et al., 2023). For example, a study by Marrone et al. (2022) investigated the relationship between AI and creativity from four key concepts: social, affective, technological, and learning factors. Results revealed that AI could certainly help them develop students’ creativity. These findings suggest that the effectiveness of generative AI in education largely hinges on students’ creativity, making their acceptance and utilization of the technology a critical factor (Li et al., 2024; Yilmaz et al., 2023). Moreover, generative AI is a powerful instrument for tackling urgent global issues related to enhancing societal wellbeing and promoting sustainable development. It is a valuable partner in tackling environmental issues such as mitigating climate change, managing resources, and developing clean energy solutions (Shafik, 2024). Also, it has economic impacts, as it is transforming industries by automating content creation, enhancing productivity, and enabling new business models. Its economic implications affect labor markets, business efficiency, and innovation (Chui et al., 2023).

The UTAUT2 model, created by Venkatesh et al. (2012), is a theoretical framework businesses use to examine influencing customer acceptance and adoption of new technology. According to this theory, behavioral intentions are influenced by four key factors: performance expectancy, social influence, effort expectancy, and facilitating conditions, which influence actual usage behavior (Venkatesh et al., 2003). Performance expectancy refers to the belief that the technology will help users accomplish tasks, aligning with perceived usefulness (PU) (Venkatesh et al., 2012). Social influence involves the degree to which individuals perceive that important others approve or disapprove of their use of the technology (Wang et al., 2009). Effort expectancy relates to how easy the technology is to use, comparable to perceived ease of use (PEOU) (Wang et al., 2009). Facilitating conditions encompass the support and resources available to users when adopting the technology (Venkatesh et al., 2012). Furthermore, Chen et al. (2020) introduced perceived enjoyment as an additional factor influencing perceived ease of use, suggesting that if users find generative AI enjoyable, they are likely to perceive it as easier to use, thereby impacting its adoption.

The present study uses the UTAUT since it is an effective framework for investigating the acceptance of generative AI, because this model predicts actual usage behavior. The authors will explore the relationships between the components of the UTAUT as applied to generative AI. The study’s novelty lies in incorporating two new variables, AI literacy and 21st century skills, into the model. Therefore, this study adds to existing literature on factors influencing the acceptance of generative AI in higher education. The study’s novelty lies in incorporating two new variables, AI literacy and 21st century skills, into the model. This study contributes to existing literature on the factors influencing the acceptance of generative AI in higher education institutions.

Ethical use is an important domain that should be investigated in order to navigate generative AI (Ray, 2023; Tan and Maravilla, 2024). Previous studies showed that it is imperative to explicitly address concerns related to fairness, accountability, transparency, bias, and integrity (Jang et al., 2022). Jang et al. (2022) identified five dimensions of AI ethics by a final instrument they developed. Five dimensions were identified: fairness, transparency, privacy, responsibility, and non-maleficence. The fairness dimension encompasses elements like considering diversity during data collection for AI development and ensuring universal disclosure of the developed AI without discrimination. The transparency dimension comprised items gauging attitudes regarding the importance of AI explainability. In the privacy dimension, there were inquiries about attitudes toward safeguarding privacy during the collection and utilization of data for creating AI. The responsibility dimension sought opinions on whether responsibilities should be allocated based on social consensus in situations involving AI-related issues or if specific groups, such as developers and users, should bear complete responsibility. Lastly, the non-maleficence dimension gathered responses on the significance of preventing abuse by various agencies associated with AI (Ryan and Stahl, 2020).

This study employed a quantitative research design to investigate the influence of AI literacy and 21ˢᵗcentury skills on college students’ acceptance of generative AI. The design utilized three validated scales to measure the key variables: AI literacy, 21st century skills, and generative AI acceptance. A cross-sectional survey was administered to collect self-reported data from participants, and statistical analyses were performed to examine the relationships between the variables.

The study was conducted at Palestine Technical University-Kadoorie (PTUK), a higher education institution in Palestine. PTUK was selected due to its diverse student population across multiple disciplines, including Information Technology, Engineering, Applied Sciences, Business, and Humanities. The university’s focus on integrating technology into education made it an ideal setting for exploring generative AI acceptance among students.

A sample of 260 college students was selected using random sampling to ensure representativeness across faculties, genders, and academic levels. The sample size was determined based on the following considerations: first, statistical Power: A sample size of 260 was deemed sufficient to achieve adequate statistical power (≥0.80) for detecting medium effect sizes in regression analyses, as recommended by Cohen (1988). Second, population Diversity: The sample included students from nine faculties, with proportional representation of males (44.6%, *n* = 116) and females (55.4%, *n* = 144), as well as Diploma (29.6%, *n* = 77), Bachelor (53.1%, *n* = 138), and Graduate (17.3%, *n* = 45) students. Third, pilot testing: A pilot study with 30 respondents was conducted to validate the adapted scales and refine the survey instrument, ensuring clarity and relevance to the Palestinian context.

This study evaluates the influence of AI literacy and 21st century skills on college students’ acceptance of generative AI. The researchers used the Likert-item rating scale. Participants self-reported their levels of AI literacy, 21st century skills, and acceptance of integrating generative AI in educational settings. To adapt the three existing scales and contextualize them to the Palestinian context, a pilot study with 30 respondents was used. Exploratory factor analysis was used for the three scales. A pre-existing AI literacy scale developed by Ng et al. (2023) was utilized and adapted. The original scale, a 32-item self-reported questionnaire on AI literacy, was developed and validated to measure students’ literacy development in the four dimensions. The adapted scale for the study using generative AI was validated, and the scale included 14 items after validation.

For assessing 21st century skills, an adapted version of the scale by Jia et al. (2016) was used in the Palestinian context. This scale originally comprised 16 items covering information literacy, collaboration, communication, innovation and creativity, and problem-solving skills. A pilot study of 30 students validated the scale in the Palestinian context. Further, the exploratory factor analysis was used to validate and adapt the scale.

A pre-existing Tugiman et al. (2023) scale was used and modified to fit the study context for generative AI acceptance. The scale was adapted by changing the context of the sentences to accept generative AI. The acceptance scale consists of five constructs: performance expectancy, effort expectancy, facilitating conditions, behavioral intention, and social influence.

Researchers employed random sampling and pilot-tested the survey with a diverse subgroup to minimize respondents’ bias. Questions used neutral phrasing and balanced scales. Respondent anonymity was ensured by omitting identifiers, using encrypted platforms, and reporting aggregated data. No names, IP addresses, or other traceable data were collected. Google Forms, which is a secure platform, was used. The study received IRB approval (No. 04/2025), and participants provided informed consent acknowledging confidentiality.

Content validity or face validity evaluates an instrument to ensure that the questionnaire has an adequate and representative group of questions that reflect the real meaning of the concept for the three pre-existing scales (Sekaran and Bougie, 2003). The respondents were asked to judge the questions’ appropriateness and suggest any items that should be included in the instrument. Pearson’s correlation for each variable assessed construct validity.

For internal consistency validity, the researchers applied the study tools to a survey sample of 30 students, and Pearson’s correlation coefficient was used for each item with the total score for the construct it relates to, and the results are shown in Table 1.

Table 1 shows that the correlation coefficients of each item with its dimension for each scale were all statistically significant at the 0.05 level, *p < 0.05, **p < 0.01, and ***p < 0.001, which confirms an appropriate degree of internal validity for the three scales.

A coefficient of at least 0.70 is required to indicate acceptable reliability (Baumgartner et al., 2002). Table 2 reports Cronbach’s alpha for each variable. Table 2 shows that Cronbach’s alpha varied from 0.70 to 0.78, which is acceptable.

Respondents in this study are from nine faculties (Faculty of Information Technology, Graduate Studies, Engineering and Technology, Applied Sciences, Business and Economics, Art and Education, and Physical Education Faculty) from Palestine Technical University- Kadoorie, comprising 116 males (44.6%) and 144 females (55.4%). There were (29.6%, n = 77) Diploma students (53.1%, n = 138). Bachelor students, and (17.3%, n = 45) Graduate Students. It also comprises 134 Scientific specializations (51.5%) and 126 Humanities specializations (48.5%). Table 3 shows the demographic characteristics.

Based on the study’s findings, it is recommended that the use of generative AI among college students be further improved by fostering critical thinking and ensuring its responsible application in education. This approach can lead to higher-quality and more effective learning outcomes. Findings also inform some recommendations for curriculum design like assessment strategies; institutions should redesign assignments to emphasize process over product, such as requiring students to document their use of AI tools and reflect on revisions. Moreover, ethical awareness by offering courses that incorporate case studies on AI-generated plagiarism and copyright dilemmas, echoing the referenced study’s call for explicit discussions about originality (Lo et al., 2025). As this study found that AI literacy and 21^st-century skills influence generative AI use, conducting qualitative studies to investigate the phenomenon deeply is thoughtful. Moreover, conducting studies to compare AI literacy rates across low-resource vs. high-resource institutions is recommended.

Some limitations could be addressed in future research. UTAUT model may not have the feature of generalizability of the findings compared to other models or theoretical frameworks. Moreover, evaluating 21st century skills poses challenges due to the lack of standardized assessment tools, which may limit the depth and accuracy of the results. Future studies could explore alternative frameworks and more comprehensive methods to address these limitations. Additionally, the tool used in this study may not have fully captured the complexity of certain constructs, such as creativity. Also, the score strength scale used in Daher (2019) lacks a clear empirical or theoretical foundation for its categorical thresholds (e.g., weak, moderate, strong). The cutoffs appear arbitrarily defined without justification, such as statistical analysis, prior validation, or expert consensus. This raises concerns about objectivity and reproducibility in this study. This arbitrariness may introduce bias, oversimplify nuanced data, and limit the scale’s validity for comparative research. Future work should employ empirically derived thresholds or continuous scoring to enhance rigor. Moreover, future studies could explore the model developed in this research by employing more detailed and objective data collection methods to measure these skills better. Another limitation is the relatively small sample size of the study, which may influence the robustness and applicability of the results.