Exploring perspectives on artificial intelligence: awareness, attitudes, and knowledge among health majors students at Saudi universities

Study type, sampling technique, sample size and participants

This study adopted a cross-sectional survey design, in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) and Survey Reporting Guideline (SURGE) recommendations. Stratified sampling was employed to ensure representation across diverse academic levels (preparatory to postgraduate) and institutional types (governmental and private). Participants were health sciences students aged 21–45 years, recruited from 12 Saudi universities, including Batterjee Medical College (Jeddah), King Abdulaziz University (Jeddah), Jeddah University (Jeddah), Umm Al Qura University (Makkah), King Saud University (Riyadh), King Faisal University (Riyadh), King Khalid University (Abha), King Fahd University of Petroleum and Minerals (Dhahran), Taibah University (Madinah), AlRayan College (Madinah), Jazan University (Jazan), and Tabuk University (Tabuk).

Questionnaire design and data collection

The questionnaire utilized in this study was designed to assess awareness, attitudes, and knowledge toward artificial intelligence among health majors’ students in Saudi universities. Students were asked to complete an online questionnaire adapted from previous studies (Al Saad et al., 2022; Hassan Mekawy, Ali Mohamed Ismail & Zayed Mohamed, 2020; Sit et al., 2020). The study utilized a questionnaire as its primary tool, which included an introductory section providing background information about the researchers and the objectives of the study, as well as a section for obtaining participant consent (Easwaran et al., 2024).

The questionnaire consisted of 22 questions organized into four sections. The first section comprised seven demographic questions addressing gender, age, nationality, university, major, academic year, and current grade point average (GPA). The second section assessed awareness of AI through questions on its definition and sources of information (e.g., academic courses, social media). The third section evaluated attitudes toward AI, exploring perceptions of its role in healthcare, its potential impact on medical careers, and the necessity of integrating AI training into curricula. The fourth section measured knowledge of AI, including understanding of core principles, types of AI (e.g., Narrow, General, Super AI), and cognitive skills related to AI programming, such as learning and reasoning.

The response options were categorized as “agree,” “disagree,” or “do not know.” A scoring system was implemented by assigning a value of “1” for each correct response and “0” for incorrect or unsure responses.

To ensure validity, the questionnaire was reviewed by experts in AI and medical education, and a pilot test was conducted. The pilot test involved 15 students (not included in the main study) to assess clarity, question flow, and estimated completion time. The average completion time was 7–9 min. Feedback led to minor revisions, including rewording two questions for clarity, adding brief explanations of AI terminology, and improving response options to reduce ambiguity. Internal consistency was confirmed using Cronbach’s alpha, yielding satisfactory reliability coefficients for awareness (α = 0.78), attitude (α = 0.81), and knowledge (α = 0.76).

The questionnaire was bilingual (English-Arabic) to accommodate participants, with the validity of the translated version confirmed through forward and backward translation.

Informed consent and ethical consideration

Ethical approval for the study was obtained from the Institutional Review Board of Batterjee Medical College, Jeddah (Reference No. RES-2023-0055). The written informed consent was obtained from all participants prior to their involvement in the study. For the online survey, participants were required to read and electronically approve the consent form before completing the questionnaire. Additionally, written and signed ethical consent forms were collected from participants as part of the ethical requirements.

Statistical analysis

The survey data were initially entered into a Microsoft Excel spreadsheet, coded appropriately, and saved as a CSV file. This file was then imported into IBM SPSS Statistics version 26 (IBM Corp., Armonk, NY, USA) for analysis. The questionnaire was structured into four main sections: (1) demographic characteristics, (2) awareness of artificial intelligence (AI), (3) attitudes toward AI, and (4) knowledge of AI. Descriptive statistics, including frequencies and percentages, were used to summarize responses for all variables and are presented in Table 1.

The data was coded aligning with practices in survey research i.e., the binary questions are coded as 1 = yes and 2 = no; if third option is included i.e., 3 = do not know. Coding for the source of information was done as 1 = academic courses, 2 = print media (books, magazines, newspapers), 3 = broadcast media (TV, radio), 4 = outdoor media (e.g., billboards, transit ads, …), 5 = social media (Facebook, Instagram, Snapchat, Twitter,…), 6 = Internet (web pages, online databases, …), 7 = video games, 8 = family members, 9 = friends and 10 = other and following question and other options in the similar manner. Chi-square ( ${\chi }^2$) tests were applied to examine associations between key categorical demographic variables (such as gender, academic year, type of university, and GPA category) and outcome variables, including levels of AI awareness (yes/no), attitudes (e.g., support for AI integration in education, willingness to learn about AI), and knowledge categories (low, moderate, high). Pearson’s chi-square test was used to identify statistically significant relationships, and the results were interpreted at three significance thresholds: p ≤ 0.05 (significant), p ≤ 0.01 (highly significant), and p ≤ 0.001 (very highly significant). These analyses aimed to highlight variations in AI-related perceptions and understanding across different student subgroups. Before performing the chi-square analysis, all categorical variables were tested for compliance with the statistical norms. For contingency table we verified that expected cell frequencies exceeded five meeting the fundamental requirement for chi-square test. This method ensured the validity of all tests of independence between categorical variables while maintaining statistical strength.

Descriptives of the participants

The investigation involved a total of 498 individuals whose demographic information is outlined in Table 1. The majority of the sample consisted of males, encompassing 328 individuals (65.9%), and 170 were females (34.1%). Analysis of the age distribution revealed that most participants belonged to the 21–25 age bracket (258 participants, 51.8%), followed by the 15–20 age bracket (200 participants, 40.2%). A smaller percentage of participants fell into the 26–30 years age groups (16 participants, 3.2%), 31–35 years (10 participants, 2.0%), 36–40 years (10 participants, 2.0%), and 41–45 years (four participants, 0.8%). The majority of participants were identified as Saudi nationals (452 participants, 90.8%), whereas individuals of non-Saudi origin constituted a minority (46 participants, 9.2%). The participants primarily hailed from public academic institutions (370 participants, 74.1%), in contrast to those from private universities (128 participants, 25.7%).

The academic disciplines that were present in the study included various fields, such as medicine (236 participants, 47.4%), pharmacy (62 participants, 12.4%), dentistry (18 participants, 3.6%), nursing (24 participants, 4.8%), radiology (10 participants, 2.0%), physical therapy (two participants, 0.4%), respiratory therapy (20 participants, 4.0%), occupational therapy (12 participants, 2.4%), healthcare administration (six participants, 1.2%), clinical nutrition (14 participants, 2.8%), medicinal laboratories (46 participants, 9.2%), and preparatory years (42 participants, 8.4%).

In terms of academic progress, the largest proportion of participants were enrolled in the preclinical year (274 participants, 55.1%), followed by the clinical year (132 participants, 26.5%), the preparatory year (50 participants, 10.0%), and the MSc or PhD level (10 participants, 4.0%). This distribution indicates that the majority of respondents were still in the early or mid-stages of their academic training. As such, their perspectives on AI may reflect foundational exposure rather than clinical experience or postgraduate specialization. Understanding their level of academic progress is important, as it may influence their awareness and attitudes toward AI depending on whether they have encountered AI applications during clinical training or coursework.

Analysis of the distribution of GPAs revealed that 34.9% of the participants had a GPA exceeding 4.5, 30.9% held a GPA ranging between 4 and 4.5, 20.9% attained a GPA between 3.5 and 4, and 13.3% obtained a GPA of 3.5 or lower.

Awareness of AI

The level of AI awareness among the participants was assessed through a series of questions, and the outcomes are delineated in Table 2.

Sources of information about AI

Participants reported diverse sources of information about AI (Fig. 1). Social media was the most common, followed by the internet, broadcast media, academic courses, family and friends, print media, video games, and outdoor advertisements.

Attitude of AI

The level of AI attitude among health major students is presented in Table 3. A substantial majority of respondents (89.2%) perceived the anticipated role of artificial intelligence in the healthcare sector as significant. Notable associations were identified between this perception and factors including university classification, academic program, and year of study (p < 0.05).

The influence of personal familiarity with AI on performance was acknowledged by a considerable majority (82.3%) of the participants. Associations of significance were established between the level of personal AI knowledge and university classification as well as academic year of study (p < 0.001 for university classification; p = 0.026 for academic year).

A considerable portion (76.7%) of the respondents expressed the belief that integrating AI education into medical training would be advantageous for medical professionals. Significant relationships were identified among gender, age, university classification, academic program, and academic year of study (p < 0.01 for all factors).

A notable percentage (60.6%) of individuals speculated about the potential replacement of certain medical specialties by AI in the coming years. Significantly, correlations were identified with age, university classification, academic program, and academic year of study (p < 0.05 for all variables).

Concerns regarding confidentiality in the utilization of AI tools for primary healthcare were shared by 67.1% of the participants. Noteworthy connections were observed among gender, age, nationality, university classification, academic program, and academic year of study (p < 0.05 for all aspects).

A substantial majority (78.3%) of respondents concurred that AI is poised to bring about a transformative revolution in the medical field within the next decade. Significant associations were documented with university classification, academic program, and grade point average (GPA) (p < 0.01 for all variables).

Approximately half (51%) of the participants believed that incorporating AI training into medical curricula should be compulsory. This viewpoint was significantly linked to gender, university classification, academic program, and academic year of study (p < 0.001 for all variables).

An overwhelming majority (78.3%) of the respondents indicated a keen interest in enhancing their understanding of AI. This enthusiasm exhibited significant associations with gender, age, nationality, university classification, and academic program (p < 0.05 for all variables).

Knowledge of AI

Table 4 presents the level of AI knowledge among students majoring in health-related fields. Approximately 43.8% of participants demonstrated an understanding of the fundamental principles of artificial intelligence. Statistically significant associations were observed between AI knowledge and gender, nationality, type of university, and academic program (p < 0.01 for all variables).

Various categories of AI based on capabilities were known by only 35.7% of the respondents. Significant relationships were identified with gender, age, type of university, program, and GPA (p < 0.001 for all).

Familiarity with the types of AI based on functions was 39.8%. Significant associations were found with sex, age, type of university, and programme (p < 0.001 for all).

A total of 49.4% of the participants were knowledgeable about AI programming-related cognitive abilities such as learning, reasoning, and self-correction. Gender, age, nationality, type of university, program, and GPA were significantly correlated (p < 0.01 for all).

The research emphasized a generally high level of awareness and positive outlook towards AI among students in the healthcare field, albeit with lower practical knowledge and participation in relevant courses. The connections observed between the demographic variables and responses indicate that factors such as gender, type of university, programme, and year of study impact students’ awareness, attitudes, and behaviours concerning AI. While many participants acknowledge AI’s potential influence on healthcare, fewer have delved deeply into the technical aspects or formal education on AI, pointing to a potential area for educational interventions. This underscores the necessity of integrating AI-related curriculum enhancements to better equip future healthcare professionals with a changing technological environment.

Demographic profile and influence on AI perceptions

The sample was predominantly composed of younger Saudi males from public universities, with a majority in the preclinical years. This demographic profile aligns with the trend of increasing technological receptivity among younger populations due to their native exposure to digital technologies from an early age (Barham et al., 2018). The associations observed between demographic variables- such as gender, type of university, and academic program- and AI-related variables highlight how these factors influence students’ engagement with AI (Pedro et al., 2019). For example, gender differences in AI awareness could be a reflection of the varying educational focus and resources allocated to males and females in different fields, potentially necessitating targeted educational interventions (Kim & Shin, 2021). These findings echo international observations, such as those by Veras et al. (2024), where demographic factors also shaped AI perceptions among health sciences students in other contexts (Veras et al., 2024).

Awareness and knowledge gaps

The study revealed a high level of awareness among Saudi university students of artificial intelligence, indicating that Saudi students are curious and interested in technology and artificial intelligence. Most of the participants (96%) defined artificial intelligence correctly, which is explained by the recent widespread spread of artificial intelligence applications and its frequent mention in the media and educational discourse due to the large amounts of data it provides, as well as the improvement of computing capabilities and the high capabilities it has demonstrated in data processing, understanding patterns and relationships, accuracy of conclusions, and quality of decision-making in specific tasks (Delello et al., 2023). Despite all this awareness among the participating students, only 28.5% had attended courses related to artificial intelligence. This low attendance reveals a critical gap between theoretical awareness and practical application, indicating the difference between what individuals know and what they can actually apply in actual work situations. This gap may be the result of several factors, including a lack of practical training and the use of traditional educational methods that focus on theory without application (Lee & Yoon, 2021; Pedro et al., 2019). Bridging this gap is crucial to enhance productivity and efficiency, as individuals capable of applying knowledge effectively contribute more significantly to organizational and healthcare goals.

For the sources of information about artificial intelligence that were adopted by the students participating in the study, social media and the internet were the dominant channels for most students. It is true that social media platforms, such as Facebook, Twitter, Instagram, Snapchat and WhatsApp, allow information to spread much faster than traditional methods do and thus help to raise awareness quickly and effectively, but they have many negative effects, such as distraction by giving rumours or biased or incorrect information (Cope, Kalantzis & Searsmith, 2021; Li et al., 2024). The use of these informal sources shows that we need to develop organized teaching plans and include them in schools. This will help make learning about artificial intelligence better and more comprehensive.

Attitudes and expectations

Attitudes towards AI were largely positive. Most respondents (89.2%) believe AI will play a pivotal role in the future of healthcare, aligning with their high expectations for AI’s contributions to clinical management (Al Saad et al., 2022). The relationships between this perception and the studied variables, such as the educational environment and the level of academic study, also indicate that students are aware of the role and importance of AI (Sit et al., 2020). This finding supports the need to improve curricula and enrich them with AI materials to keep pace with students’ perceptions and ensure their readiness for the profession of the future.

The positive attitude towards AI continues, as the majority of participants (82.3%) acknowledged the impact of personal familiarity with AI on performance, reflecting Saudi students’ awareness of the impact of personal experience on the quality of what a person can offer and how this person can exploit and adapt artificial education technologies in a beneficial way in the healthcare sector (Ali Mohamad et al., 2023). We found that this perception increases with increasing academic level and may differ from one university to another, which confirms the importance of education in shaping and formulating students’ understanding of and competence in the use of AI technologies and programs, which will be positively reflected in public health later (Ahmad et al., 2023).

Advocating for AI education would be beneficial for medical professionals, as 76.7% of male and female students from all academic programs believe. This indicates a clear demand and desire to learn AI and train in its applications and programs (Fazil et al., 2024), as it will have a clear impact on increasing the level of health professions and effectively using technology to meet diverse needs (Cheng et al., 2020).

Despite all the previous strong positive trends, we found that only 60.6% of the participating students expected some medical specialties to be replaced by AI in the coming years, with some differences according to age, university classification, academic program, and year of study. Additionally, 67.1% of the participants were concerned about confidentiality in the use of AI tools in primary health care, with some differences between different demographic and academic factors. These results indicate students’ concerns about their professional future and whether their roles will be replaced by programmed machines (Li et al., 2022). Hence, the importance and role of the educational institution in addressing these concerns through educational curricula will focus on the importance of building personal knowledge and ethical considerations and integrating them in the field of AI and how to benefit from its tools and applications (Akgun & Greenhow, 2022), ensuring that healthcare professionals are able to effectively manage privacy in their future careers (Akgun & Greenhow, 2022; Pedro et al., 2019).

Even with all the concerns that AI threatens students’ professional future, 78.3% of them believe that AI is poised to transform the medical field better over the next decade, with little variation in this belief depending on the rankings of universities, academic programs, and students’ GPA. This belief among students highlights their belief in the great potential of AI in healthcare (Bohr & Memarzadeh, 2020). This requires the development of customized educational strategies that prepare students for increasing technological advancements and enhance the integration of AI into the healthcare landscape (Osunlaja et al., 2024).

One surprising finding is that despite most students’ preference for AI, half of them do not believe that they should receive AI training as part of their curriculum. This suggests a significant divide in perceptions regarding the importance of AI in their education. This may be explained by students feeling that they do not need AI immediately and that their current curriculum is sufficient to prepare them to be competent professionals (Teng et al., 2022). To address this divide, we reiterate the importance of integrating AI into the curriculum and better highlight the value of using AI tools and techniques and their potential impact on their future professional careers (Chiu & Chai, 2020).

Among the attitudes toward AI, 78.3% of the health science students who participated in the study expressed a strong interest in furthering their understanding of AI, which explains their strong desire to engage more deeply with this technology. This is explained by the fact that students are always motivated by curiosity, independence, and self-learning, and the use of AI encourages them to explore educational materials further and improve academic performance by helping them perform better on tests and exams through the provision of personalized reviews and learning materials (Chen, Chen & Lin, 2020; Huang, Lu & Yang, 2023). The study also indicated that this belief varied across demographic and educational backgrounds, reflecting differences in initial exposure to AI, the perceived importance of AI for specific medical specialties, or the influence of institutional attitudes toward the technology (Stewart, Rybicki & Dwivedi, 2020). These data emphasize the importance of everyone having equal opportunities to learn about AI and prepare to use these technologies effectively in their future careers (Bohr & Memarzadeh, 2020).

Knowledge and technical competence

Despite strong awareness and positive attitudes, students’ technical knowledge was moderate. Only 43.8% reported understanding AI’s basic principles, with significant differences across demographic and academic factors. The result is logical given that AI is a new science worldwide and has only recently been included in educational curricula in recent years (Stolpe & Hallström, 2024). This moderate knowledge of the basics of this science can be attributed to differences in curricular focus or educational resources that affect the possibility of teaching AI in different universities and educational institutions (Chen, Chen & Lin, 2020; Smit, Zoet & van Meerten, 2020).

The study also highlighted AI types, which are classified into two categories: first, capabilities (narrow AI (dedicated to a single task), general AI (working like a human), and super-AI (smarter than humans)) and second, functions (interactive machines (machine learning), limited memory (deep learning), theory of mind (emotions and interactions), and self-awareness (future AI with self-awareness)). When health college students were asked about both types of AI, only 35.7% of the participants were aware of capability-based AI typologies, and 39.8% of the participants were aware of function-based AI typologies. This lack of knowledge reveals a significant gap in students’ understanding of the potential of diverse AI technologies, highlighting the need for more comprehensive AI education that effectively covers these fundamental distinctions (Al Saad et al., 2022; Saghiri et al., 2022). Significant associations with demographic and academic factors also suggest that students with stronger academic performance or specific educational experiences may have a better understanding, suggesting that educational backgrounds and demographic factors influence knowledge. This suggests that we need to educate students in general and health professionals in particular about how AI works, regardless of where they come from or their background (Chung, Thaichon & Quach, 2022; Cope, Kalantzis & Searsmith, 2021).

Regarding cognitive AI programming skills such as learning, reasoning, and self-correction, nearly half of the participants (49.4%) reported familiarity. The study further found that students with higher academic performance or those enrolled in technologically advanced universities demonstrated a better understanding of these skills. These findings highlight the importance of developing targeted educational interventions to enhance all students’ comprehension of advanced AI concepts (Cope, Kalantzis & Searsmith, 2021; Khanagar et al., 2021; Shin, 2020).

After the knowledge of students is studied, a critical gap in AI education can be identified; therefore, it is necessary for universities and colleges to implement comprehensive training in the field of AI that covers not only the basic principles but also the capabilities and functions of more complex AI. This is important because we want to prepare them for jobs where they can use intelligent computer technology, such as assisting doctors, nurses, technicians in hospitals and clinics and many other jobs.

Limitations

The current study was designed and performed to assess the awareness, knowledge, and attitudes of study participants regarding AI, with the intention of informing curriculum design and policy development for future learners. However, several limitations should be considered. The cross-sectional design limits the ability to establish causality and may have induced selection bias, potentially leading to the overrepresentation of certain demographics, such as younger males from public universities. Additionally, reliance on self-reported data could introduce response bias, as participants may overestimate their actual AI competence. The study also did not include follow-up assessments, limiting the evaluation of knowledge retention or skill development over time. Furthermore, complexity related to potential confounding variables may have influenced the observed associations. Future research should incorporate longitudinal or interventional designs—such as following a cohort through AI-integrated curricula—to better establish causal relationships and validate findings. Finally, further questions assessing AI literacy and the integration of cognitive skills need careful framing to ensure clarity. A pilot study is recommended, which could involve either testing the consistency and validity of the refined questionnaire or evaluating the feasibility and impact of educational interventions.