This study employed an exploratory, descriptive cross-sectional survey design and was carried out during the Spring 2025 semester with students and alumni from the College of Health Sciences, Dental Hygiene Department at the University of Doha for Science and Technology (UDST). The study was conducted via an online survey utilizing the SurveyMonkey platform¹ and was circulated online via email to dental hygiene students and alumni. Based on similar studies without any prespecified hypothesis, the sample size estimation was waived (Alam et al., 2024; Eschert et al., 2022; Roganović et al., 2023; Yılmaz et al., 2024). This study was explorative and did not begin with a predetermined hypothesis; therefore, calculations for sample size and power estimation were not required, consistent with prior exploratory survey studies in this area. Nevertheless, the absence of a priori sample size calculation may limit statistical power for detecting smaller associations and may reduce the precision of estimates, which should be considered particularly for subgroup analyses, while the single-institution sampling design constrains external validity. Accordingly, subgroup summaries (e.g., gender, student’s vs. alumni, and across academic years) are presented descriptively to characterize response patterns rather than to support inferential conclusions. Any apparent subgroup differences should be interpreted with caution. A convenience sampling method was employed, which involved selecting participants who were easily accessible and willing to participate. The study’s methodology was designed in accordance with the Checklist for Reporting of Survey Studies (CROSS), a consensus-based guideline developed to enhance the transparency, consistency, and quality of survey research reporting. The checklist includes key elements such as clear articulation of study design, detailed description of the survey instrument, sampling methods, data collection procedures, and ethical considerations (Sharma et al., 2021).

Before commencing the study, approval was obtained from the HBKU Institutional Review Board (IRB #: HBKU-IRB-2025-137), ensuring all research protocols meet the required ethical standards. To safeguard participant confidentiality and voluntary participation, electronic informed consent was obtained from all participants. This consent process clearly communicated the study’s purpose, procedures, and potential benefits, allowing participants to make an informed decision about their involvement. Additionally, all data collected were anonymized and stored securely to prevent unauthorized access, thereby maintaining the confidentiality of participant information. The study posed no risks to participants and did not collect identifiable information. Participation was entirely voluntary, and participants could withdraw at any point without consequence. While there were no direct personal benefits from participating, the information provided was invaluable in enhancing the understanding of the educational needs surrounding AI in dental hygiene. This can led to significant improvements in the curriculum, ultimately benefiting future dental professionals by better preparing them for the integration of AI technologies in their practice.

A 29-item questionnaire, detailed in Appendix A, was utilized, adapted from the validated survey designs (Hammoudi Halat et al., 2024; Sindermann et al., 2021). The survey consists of five distinct sections, each targeting specific study aspects. The first, second, fourth, and fifth sections of the questionnaire were adopted from Hammoudi Halat et al. (2024) which reported a high internal consistency (Cronbach’s α = 0.93). Hammoudi Halat et al. (2024) was conducted on dentistry students, making it particularly relevant to the present study’s population of dental hygiene students and alumni. The third section focused on attitudes toward artificial intelligence employing the validated Attitude Towards Artificial Intelligence Scale (ATAI) adapted from Sindermann et al. (2021), which included university students, with Cronbach’s α values ranging from 0.61 to 0.73 across different cultural samples. Although only five items were included in this survey, which is considered a low number, these reliability coefficients were deemed acceptable. Both sources validated their respective scales, ensuring the robustness of the survey. The survey questions were constructed based on the following five main sections, each comprising multiple measures designed to capture distinct aspects of the study (Figure 1).

Recruitment occurred through institutional email distribution by UDST to dental hygiene students and alumni. A total of 119 eligible individuals (54 students and 65 alumni) were invited to participate between February and March 2025. The email described the study aim and eligibility criteria, included a SurveyMonkey link, and encouraged recipients to answer the survey. To participate in the study, the dental hygiene students had to be enrolled in or previously graduated from UDST. Additionally, two reminders were sent following the initial email to maximize participation. Informed consent was secured through the online form at the beginning of the survey, and a download link for the PDF-approved soft copy was provided at the beginning of the survey for the participant’s reference. SurveyMonkey settings were configured to prevent duplicate responses by restricting participants to a single submission, ensuring that each participant could submit only one response. The submission of the survey was taken as consent to participate in the study.

Data collected from the survey were systematically exported to an Excel spreadsheet, which was meticulously cleaned to ensure accuracy and consistency, thereby creating a reliable database for analysis. Additionally, the cleansed data were analyzed using the advanced statistical capabilities of R-Studio, version 4.4.2, and Microsoft Excel 4, version 2,505. The statistical methods of calculating continuous variables, categorical variables and scores of the survey were adopted from earlier studies (Hammoudi Halat et al., 2024; Sindermann et al., 2021) which are the original studies where the survey instruments were first developed and utilized. Although the analyses were exploratory and not hypothesis-driven, standard regression diagnostics were performed to ensure the validity and interpretability of the models. Before conducting linear regression analyses, standard model assumptions were assessed. Linearity and homoscedasticity were evaluated through visual inspection of residual-versus-fitted and scale-location plots, while normality of residuals was assessed using Q-Q plots. Potential influential observations were examined using residuals-versus-leverage plots. No violations of regression assumptions were identified, supporting the appropriateness of linear regression for the data.

A total of 119 invitations were sent, yielding 101 responses for an approximate response rate of 84.87%, which is considered a highly representative capturing the majority of the target population. 54 (53.5%) were dental hygiene students and 47 (46.5%) were alumni. The age range of participants spanned from 18 to 38 years, with a male to female ratio of 1:8. Demographic details of the participants are summarized in Table 1.

In assessing the baseline AI knowledge among participants, the findings reveal a varied baseline understanding (see Figure 2). A minority of the participants, constituting 13.9% (n = 14), reported having no prior knowledge of AI, underscoring a gap in fundamental AI awareness within this group. The majority, however, demonstrated some level of familiarity, with 77.2% (n = 78) acknowledging only basic AI knowledge. This suggests that while AI concepts might be recognized, the depth of understanding is potentially superficial. Notably, a small segment, 8.9% (n = 9), claimed advanced knowledge, indicating a disparate level of expertise that could influence perceptions and attitudes to integrate AI technologies into professional practices. Furthermore, a significant discrepancy in the educational background of participants regarding AI in healthcare was noted, with a notable majority, 68.3% (n = 69), reporting no prior educational or training experience in AI, suggesting a foundational gap in AI knowledge among participants. For those who had engaged in some form of AI education, the experiences were diverse but limited in reach: 10.9% (n = 11) participated in university-based in-person courses, 3.96% (n = 4) in university-based online courses, 7.9% (n = 8) in non-university-based courses, and 8.9% (n = 9) in other educational activities. Moreover, the effectiveness of these educational initiatives as perceived by participants varied: 7.9% (n = 8) rated their AI training as having little usefulness, and 17.8% (n = 18) as of average usefulness, which may indicate inadequacies in the AI educational training they received. Conversely, a more significant proportion found the training to be beneficial, with 38.6% (n = 39) rating it as very useful and 14.9% (n = 15) as extremely useful, underscoring the potential impact of well-structured AI training programs. Interestingly, 20.8% (n = 21) considered these experiences not applicable, possibly reflecting either the irrelevance of the training to their professional roles or an absence of formal AI education.

To present an overview of participants’ responses regarding AI acceptance and fear, the ATAI scores were analyzed based on gender and education level (Table 2). The overall Acceptance score among the dental hygiene participants was 6.56, indicating that, on average, they Slightly Agreed with statements reflecting trust in and perceived benefits of AI. On the Fear scale, participants scored an average of 4.85, suggesting a stance between Slightly Disagreeing and Neither Agree nor Disagree regarding concerns about AI’s potential risks. Moreover, the descriptive analysis of AI attitudes by gender and education level highlights variations in reported acceptance and fear of AI across participant groups. While both male and female participants exhibited similar levels of AI acceptance, with mean scores falling within the ‘Slightly Agree’ to ‘Mostly Agree’ range (Males = 6.86; Females = 6.53). However, a difference was noted in the reported fear levels. Male participants recorded a lower mean fear score (3.48), corresponding to the ‘Mostly Disagree’ to ‘Slightly Disagree’ range, whereas female participants reported a higher mean score (5.02), aligning with the ‘Neither Agree nor Disagree’ category’. Across education levels, first-year students demonstrated the highest AI acceptance and the lowest fear, whereas second-year students reported lower acceptance and higher concern about AI-related risks. Third- and fourth-year students reported moderate responses across both dimensions, with fourth-year students exhibiting a slightly higher acceptance of AI. Alumni maintained moderate attitudes regarding both AI acceptance and fear. These findings offer insights into how AI attitude was distributed within the sample, though variations in group sizes should be considered when interpreting the results as an exploration of response patterns rather than definitive trends or basis for inferential conclusions. The small sample size limits the generalizability of these findings. Subgroup analyses were descriptive, and inferential conclusions should be interpreted with caution.

The results of perceptions of AI in health education (Table 3) suggest that participants generally hold a neutral to positive perception regarding the incorporation of AI in health education. A considerable portion of participants (69.3%) either agreed or strongly agreed that integrating AI into health education would ease the learning process (M = 2.50, SD = 1.37). Similarly, over half of the students (52.5%) expressed willingness to use and benefit from AI during their university education (M = 2.31, SD = 1.45). The inclusion of AI in health education was also well-received, with 70.3% supporting its incorporation into university education (M = 2.29, SD = 1.34). However, there was some uncertainty regarding AI’s role in real-life practice, as 41.6% of students responded neutrally when asked if AI would prepare them for future professional applications (M = 2.55, SD = 1.21). Despite this, most students remained optimistic about gaining a better understanding of AI in healthcare by the end of their degree, with 69.3% either agreeing or strongly agreeing (M = 2.26, SD = 1.30). The overall mean perception score of AI in health education (M = 2.38, SD = 1.33) suggests that while students recognize the potential benefits of AI in health education, some remain neutral or uncertain about its practical implications.

For the results of AI topics to be included in health education (Table 4), AI for health-related research received the highest importance rating (M = 3.33, SD = 0.69), with the majority (50.5%) considering it important and 42.57% rating it as very important. Similarly, AI in mobile health applications for patient support was highly rated (M = 3.20, SD = 0.70), with 49.5% rating it as important and 35.64% as very important. AI in surveillance and epidemic control (M = 3.10, SD = 0.67) and AI for reducing errors in healthcare (M = 3.10, SD = 0.74) also received high importance ratings, with the majority of respondents considering them either important or very important. AI in radiology and imaging procedures (M = 3.04, SD = 0.72) and AI in disease prevention (M = 3.03, SD = 0.78) were also considered relevant, with most respondents rating them as important or very important. Moreover, knowledge and skills related to AI applications in healthcare (M = 2.98, SD = 0.72) and AI in new drug development (M = 2.90, SD = 0.79) were regarded as important by the majority of respondents. Topics such as AI in genetics and genomics (M = 2.81, SD = 0.83), AI in disease monitoring (M = 2.82, SD = 0.81), AI in disease diagnosis (M = 2.71, SD = 0.81), and AI and robotics in surgery (M = 2.77, SD = 0.95) received moderate importance ratings, indicating a varying level of perceived relevance among participants. Overall, these findings highlight the strong recognition of AI’s role in various healthcare domains, with the highest importance placed on AI applications in research, mobile health, and epidemic control.

The findings indicate that AI perception in health education significantly influences AI acceptance (R² = 10.66%, p = 0.0008613), albeit with a weak-to-moderate positive correlation (r = 0.3265), suggesting that while perceptions of AI in health education play a role, other factors exert a stronger influence. Similarly, the perceived usefulness of previous AI training significantly impacts AI acceptance (R² = 14.59%, p = 0.004123), where lower-rated usefulness corresponds to lower acceptance, highlighting the importance of training quality and its perceived relevance in shaping AI adoption. In contrast, variables such as age (R² = 0.028%, p = 0.8681), gender (R² = 0.31%, p = 0.5819), and education level (R² = 6.8%, p = 0.1448) Were found to have no statistically significant impact on AI acceptance. Additionally, perceived AI knowledge revealed a statistically significant but weak relationship with AI acceptance (R² = 7.53%, p = 0.02156), with individuals reporting no AI knowledge demonstrating lower acceptance. On the other hand, previous experience or training in AI showed no significant impact on AI acceptance (R² = 1.99%, p = 0.7441), indicating that merely having exposure to AI-related education or training does not necessarily translate to higher acceptance levels.

In both models of multiple linear regression analyses, perceptions of AI in health education emerged as the main significant predictor (Model 1: β = 0.660, p = 0.0137; Model 2: β = 0.676, p = 0.007), indicating that individuals with more favorable views toward AI were more likely to accept AI. Demographic variables including gender, age, and year of study as well as self-evaluated AI knowledge and prior training showed no significant associations with AI acceptance. However, in the second model, the perception that AI training for the “Not applicable” responses was significantly associated with lower AI acceptance (β = −1.570, p = 0.017). Model 1 (Table 5) explained 28.2% of the variance in AI acceptance (R² = 0.2821, adjusted R² = 0.1351), while Model 2 (Table 6) explained 26.0% of the variance (R² = 0.2602, adjusted R² = 0.1688). The slight increase in adjusted R² from 0.1351 to 0.1688 in the second model suggests that demographic variables did not add substantial explanatory power to the model. The demographic factors removal did not alter the overall conclusions, suggesting that AI acceptance is mainly driven by perceptions rather than demographics. The F-statistic (F = 2.846, p = 0.003) confirmed the overall statistical significance of the second model, indicating that despite the limited number of significant individual predictors, the model as a whole provides valuable insights into AI acceptance.

The findings indicate that while the majority of participants (77.2%) had basic AI knowledge, only a small proportion (8.9%) reported advanced AI knowledge. Furthermore, 68.3% had no formal AI education, emphasizing a substantial gap in AI literacy. These results align with prior studies indicating that healthcare students and professionals often have limited formal training in AI, despite growing AI applications in clinical practice (Allam et al., 2024; Aboalshamat, 2022; Al-Qerem et al., 2023; Wood et al., 2021; Weidener and Fischer, 2024). However, this study expands on previous research by illustrating that even those with AI exposure report variability in the perceived usefulness of their dental training. While 38.6% found AI training very useful, a notable portion (20.8%) considered it “Not applicable” raising concerns about the relevance and effectiveness of AI curricula. The “Not applicable” response may indicate several possibilities: A lack of engagement in AI training due to perceived irrelevance to their field, unawareness of available AI training opportunities, incomplete or inadequate AI training experiences that failed to provide meaningful insights, or skepticism about AI’s role in dental hygiene, leading participants to dismiss AI training as unnecessary. The ‘Not applicable’ responses contrast with findings from previous studies in medical education, where the integration of AI in educational courses has been proven useful and shown to enhance the teaching and learning experience (Naseer et al., 2025; Yang and Shulruf, 2019). The lack of formalized AI education in dental hygiene programs may contribute to skepticism or uncertainty regarding AI’s practical benefits. This highlights the necessity for AI-focused curriculum reforms that not only introduce fundamental AI concepts but also provide relevant and hands-on experience in AI applications specific to oral healthcare.

Overall, dental hygiene participants’ perception score expressed that while participants recognize the potential benefits of AI in health education, some remain neutral or uncertain about its practical implications (M = 2.38, SD = 1.33). Furthermore, uncertainty emerged regarding AI’s applicability to real-world practice, with 41.6% of participants responding neutrally when asked if AI would prepare them for professional applications. This hesitation parallels findings in other healthcare disciplines, where pharmacy students acknowledge AI’s potential but question its direct impact on their critical thinking and judgment skills (Knobloch et al., 2024). Additionally, a considerable portion of dental hygiene participants agreed that AI integration would ease learning (69.3%) and were willing to use AI in their education (52.5%). These results align with prior studies suggesting that dentistry students generally recognize AI’s educational benefits, evidenced by 73% of the students agreeing that AI should be included in their health education (Hammoudi Halat et al., 2024).

Dental hygiene participants rated AI for health-related research (M = 3.33, SD = 0.69) as the most important AI topic. These findings align with research indicating that dentistry students rank AI applications that enhance research as the second most important priority (Hammoudi Halat et al., 2024). AI in mobile health applications, surveillance, epidemic control, and error reduction in healthcare also received high ratings from the dental hygiene students, reflecting their view of the broader recognition of AI’s role in improving healthcare efficiency. However, topics such as AI in surgery (M = 2.77, SD = 0.95) and AI in disease diagnosis (M = 2.71, SD = 0.81) were rated lower, suggesting that dental hygiene students perceive these areas as less relevant to their profession. This highlights the need to tailor AI education to discipline-specific applications rather than adopting a one-size-fits-all approach.

The regression analysis identified AI perception in health education as the strongest predictor of AI acceptance (β = 0.676, p = 0.007), aligning with previous study demonstrating that dental students with positive AI perceptions are more likely to be ready to adopt AI technologies (Hammoudi Halat et al., 2024). In contrast, age, gender, education level, and prior AI training showed no significant correlation to AI acceptance. Similarly, Hammoudi Halat et al. (2024) reported that age, gender, education level, and prior AI training have no significant correlation with AI readiness. Interestingly, dental hygiene participants who responded to AI training usefulness as “Not applicable” exhibited significantly lower AI acceptance (β = −1.570, p = 0.017). This suggests that those who do not engage with AI education or perceive it as irrelevant are less likely to embrace AI applications. Therefore, we advocate for the development of structured AI curricula that can lead future dental hygienists to perceive and accept AI.

The present study identified moderate AI acceptance, neutral-to-slight fear, and low-to-basic knowledge, with perception of AI in health education emerging as the strongest predictor of acceptance, a pattern that aligns with international evidence while also revealing contextual differences. In South Korea, high interest in AI was coupled with professional caution, low confidence in diagnostic use, and a strong demand for AI education (Jeong et al., 2024; Ezzeldin et al., 2025), reflecting the same concern between interest and clinical trust observed in the present cohort. Egyptian dental students likewise demonstrated positive knowledge, perceptions, and attitudes toward AI, particularly for CAD/CAM and implant applications, while rejecting the idea that AI could replace dentists and expressing strong interest in further AI education (Ezzeldin et al., 2025), consistent with the present participants’ preference for AI applications in health-related research and mobile health, and their view of AI as a supportive rather than a replacement technology. In Australia, AI was widely viewed as beneficial but accompanied by uncertainty and concerns about accuracy and job displacement, with attitudes influenced by some demographic factors (Hegde et al., 2025), whereas no such demographic effects were observed in the present study, suggesting that perceived relevance and curricular exposure may be more influential in the present study. Similarly, Saudi patients’ acceptance of AI was driven primarily by perceived benefits rather than by demographic characteristics or perceived risks (Sharka et al., 2025), mirroring the present regression findings in which perception was the strongest predictor of acceptance. Collectively, these international findings reinforce the importance of practical, experience-based AI education in dental training (Jeong et al., 2024; Ezzeldin et al., 2025; Hegde et al., 2025; Jeong et al., 2023; Sharka et al., 2025).

While this study offers valuable insights into the attitudes, perceptions, and knowledge of AI among dental hygiene students and alumni, several limitations should be acknowledged when interpreting its findings, as they also present opportunities for future research. The study employed a cross-sectional survey design, capturing data at a single point in time, which limits its ability to account for changes in AI perceptions over time; a longitudinal approach would allow for tracking the evolution of attitudes as students progress through their education and transition into professional practice, offering a more dynamic understanding of AI acceptance. In addition, because the sample included both students and alumni, differences in post-graduation exposure to digital and AI-based technologies may have influenced responses. Although education status (student vs. alumni) was included in the regression model and was not a significant predictor of AI acceptance, unmeasured differences in professional maturation, professional experience, workplace technology adoption, and informal learning after graduation may still have introduced residual confounding. Future studies should therefore use stratified or longitudinal designs to more precisely isolate the effect of professional experience on AI attitudes. Furthermore, although the study achieved a high response rate (84.87%) from the dental hygiene participants at UDST, the sample was limited to a single institution in Qatar, which restricts generalizability to other regions, dental disciplines, or educational systems. Additionally, convenience sampling and the absence of a priori sample size calculation may limit generalizability and reduce the ability to detect small but potentially meaningful differences between participant subgroups (e.g., by gender or education level). Broader studies incorporating larger, more diverse samples across multiple institutions and countries are needed to enhance external validity. In particular, future hypothesis-driven studies with adequate sample sizes and multi-institutional designs should formally evaluate differences across academic years using appropriate inferential and post-hoc analyses. Another notable limitation is the gender imbalance within the sample, as 89.1% of participants were female, reflecting the female-dominated nature of the dental hygiene profession (Luciak-Donsberger, 2003; Kiser, 2022; Grenestedt and Donk, 2021; Perri, 2024). While gender was recorded and examined descriptively and analytically, it was not treated as a primary explanatory variable, as knowledge acquisition and professional attitudes in dental hygiene are largely shaped by standardized academic curricula and clinical exposure rather than gender-based differences. Nevertheless, imbalance may have influenced findings related to gender differences in AI acceptance and fear levels; future research with more balanced gender representation would allow for deeper exploration of these gender-related effects. The reliance on self-reported data introduces potential biases, including social desirability bias, recall bias, and misinterpretation of AI-related terminology, which may have led participants to inaccurately assess their AI knowledge or attitudes. Furthermore, the study excluded dental hygiene faculty members, dental educators, and practicing dentists, whose perspectives on AI integration in curricula and clinical practice could offer critical insights into real-world challenges and opportunities. Future research should aim to include these groups to provide a holistic view of AI adoption in dentistry. Despite these limitations, this study establishes important baseline data on AI-related attitudes and knowledge among dental hygiene students and alumni in this part of the world. Addressing these limitations in future studies will help fill existing knowledge gaps, refine AI training strategies, and explore adoption patterns across diverse dental specializations and professional levels, ultimately contributing to a more comprehensive understanding of AI integration in dental education and clinical practice.

The study highlights key considerations for incorporating AI into dental hygiene education: