Abstract
Background
This cross-sectional study investigates healthcare practitioners’ perceptions of the implementation of artificial intelligence (AI) to enhance patient safety culture in Riyadh, Saudi Arabia, conducted from March to June 2025. The focus is on understanding how AI is perceived in the context of improving patient safety culture and aligning with Saudi Vision 2030 goals.
Methods
The study employed a cross-sectional design and administered self-administered online surveys via convenience sampling among healthcare practitioners across multiple healthcare settings in Riyadh. The study targeted a population including doctors, nurses, and allied health professionals. A structured questionnaire was developed to assess perceptions of AI, including key variables such as perceived AI benefits, concerns about data privacy, and the necessity of training. Descriptive statistics were computed to characterize the sample, including age, gender, and profession.
Results
Findings revealed that most participants perceived AI as supportive of patient safety through improved diagnostic accuracy, reduced medical errors, and streamlined workflows, which participants believed may contribute to a stronger patient safety culture. Descriptive analyses suggested variation in perceptions across professional groups.
Discussion
The study found that healthcare practitioners in Riyadh generally perceived AI as a potentially valuable tool for supporting patient safety culture, particularly through improving diagnostic accuracy and reducing errors; however, concerns about data privacy and insufficient training remain significant barriers that must be addressed to ensure effective and safe AI integration.
Conclusions
This study highlighted healthcare practitioners’ positive perception of AI’s role in enhancing patient safety culture in Riyadh. While AI is seen as beneficial in improving accuracy and reducing errors, challenges such as data privacy concerns and a lack of training were identified as barriers to its implementation.
Clinical trial number
Not applicable.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12913-026-14901-8.
Keywords: Artificial intelligence, Patient safety, Safety culture, Saudi Arabia, Healthcare practitioners
Introduction
Artificial intelligence (AI) is transforming healthcare in developed countries and has the potential to significantly advance global health [1, 2]. AI-driven technologies are being integrated into various aspects of the healthcare system, including diagnosis, treatment planning, and medication discovery, thereby enhancing overall accessibility, efficiency, and quality of care [3, 4]. The deployment of AI tools may support healthcare practitioners through reducing human error, improving diagnostic accuracy, and facilitating communication and collaboration., thereby fostering a stronger safety culture [5, 6]. AI involves developing systems that replicate human cognitive functions, such as learning and problem-solving, using algorithms. These systems are adaptive, intelligent, and capable of analyzing complex datasets, making them effective for tasks such as interpreting medical data. AI is also dynamic and self-improving as it processes new data [7, 8].
Patient safety refers to preventing errors and adverse outcomes in healthcare delivery. It is a critical issue in the healthcare sector [9, 10]. It aims to reduce the incidence of adverse outcomes and medical errors and preventable patient harm. Healthcare professionals can mitigate the risks of complications, injuries, and deaths by enforcing regulations and safety protocols in their practice. Hospital systems are recognized as the primary environments for mitigating medical errors and improving the quality of patient care. Optimal patient outcomes result from safe treatment. Safe healthcare practices are associated with improved patient outcomes, faster recovery, fewer complications, and fewer hospitalizations when they receive high-quality, safe care [11]. AI has the capacity to revolutionize healthcare delivery by enhancing diagnostic precision, forecasting patient outcomes, and optimizing clinical operations [12]. The degree to which AI can improve the patient safety culture among healthcare professionals in Saudi Arabia remains inadequately investigated [12].
Saudi Arabia is undergoing a significant transition in healthcare, with the integration of AI aimed at enhancing patient safety culture among healthcare practitioners. This transition aligns with the country’s Vision 2030, which seeks to modernize the healthcare system by integrating advanced technologies to enhance service delivery and improve patient outcomes [1]. Initiatives under the Ministry of Health’s National Transformation Program focus on developing AI capabilities and providing necessary training to healthcare professionals to ensure seamless integration. However, challenges such as data privacy concerns, the need for robust training programs, and the ethical use of AI require attention [3]. Addressing these challenges is crucial to leveraging AI to enhance patient care and strengthen the safety culture among practitioners, ensuring AI’s full potential for improving patient safety is realized [3].
Conceptually, the integration of artificial intelligence into healthcare can be understood through established frameworks of technology adoption and safety culture. Models such as the Technology Acceptance Model emphasize how perceived usefulness and ease of use influence practitioners’ engagement with digital tools, while patient safety culture frameworks highlight organizational learning, communication, and error prevention. AI systems that support clinical decision-making, monitoring, and workflow coordination may reinforce these dimensions of safety culture by reducing cognitive burden, improving information accuracy, and facilitating timely interventions. Thus, healthcare practitioners’ perceptions of AI adoption are closely linked to their perceptions of safety culture, providing a theoretical basis for examining these constructs together.
Several key gaps hinder the successful implementation of AI in healthcare settings. A significant gap in the effective use of AI in healthcare is the lack of professional training for healthcare practitioners. Many healthcare workers lack the knowledge and skills to understand, evaluate, and effectively use AI solutions. This may result in distrust, underutilization, or potential abuse of AI technology. Inadequate training may prevent healthcare practitioners from realizing the full potential of AI, resulting in missed opportunities for improved patient care [13]. Another significant difficulty is the quality and accessibility of healthcare data. AI systems rely heavily on high-quality, extensive datasets to achieve precise and dependable outcomes. However, healthcare data often reveals inaccuracies, biases, and fragmentation. This may contribute to inaccurate AI predictions, potentially leading to misdiagnosis, inappropriate or suboptimal treatment options, and even patient harm [12]. Furthermore, ethical and legal issues substantially restrict the successful implementation of AI in healthcare. AI algorithms trained on historical data may inherit biases, leading to biased outcomes and exacerbating health inequities. Data privacy and security are essential and require stringent procedures to safeguard sensitive patient information [14].
This study aims to investigate the current implementation status, prevailing beliefs, and perceived challenges associated with the adoption of AI technologies in healthcare settings in Riyadh, Saudi Arabia. By generating localized empirical evidence, our research aims to address a significant gap in the literature regarding AI integration and its implications for patient safety culture, particularly in the context of Saudi Vision 2030. Although many international studies have examined AI applications in healthcare, few have focused on their contextual relevance and impact within the Saudi healthcare system [5]. The incorporation of AI into healthcare practices can significantly reduce medical errors, enhance diagnostic precision, and optimize clinical workflows, ultimately improving patient safety and enhancing system resilience. The findings of this study are expected to provide evidence-based recommendations for healthcare administrators, policymakers, and technology developers to address implementation barriers and capitalize on opportunities for enhancing AI-driven tools in Riyadh’s healthcare sector, ultimately leading to improved patient outcomes and a more robust safety culture [3].
Methods
Study design, setting, participants
A cross-sectional, self-administered, online Survey using convenience sampling among healthcare practitioners (HCPs) in Riyadh, Saudi Arabia. Participants eligible for the study were physicians, nurses, pharmacists, and administrative personnel working in healthcare facilities in Riyadh, Saudi Arabia. Individuals not directly involved in the delivery or management of healthcare were excluded. Participants were recruited through electronic invitations sent via institutional email lists, hospital intranet portals, and WhatsApp groups aimed at healthcare practitioners in both the public and private sectors. The survey link was distributed through associations and social media groups frequently used by healthcare professionals to increase reach and diversity. The data collection occurred from March to June 2025. Participation was entirely voluntary, and informed consent was secured electronically before accessing the questionnaire. Personal identifiers were not collected, and participants could withdraw at any time without penalty. Respondents were not offered any financial or material incentives.
Instrument
Data were gathered using a self-administered, structured questionnaire developed by the research team after a thorough review of the pertinent literature on artificial intelligence in healthcare and patient safety culture [12, 13]. The questionnaire items were developed through a structured review of published literature addressing artificial intelligence adoption in healthcare and patient safety culture. Previously reported survey instruments and conceptual domains from related studies informed the initial item pool to ensure content relevance and construct alignment. Draft items were generated to capture practitioner perceptions of AI implementation, perceived benefits, barriers, and safety culture dimensions. These items were then evaluated by a multidisciplinary expert panel comprising specialists in health informatics, patient safety, and public health to assess their clarity, relevance, and contextual suitability. Feedback from this review guided the refinement of items before pilot testing. The pilot phase focused on evaluating item clarity, comprehension, and internal consistency rather than scale validation, resulting in minor wording adjustments prior to final deployment. The instrument consisted of four primary domains encompassing a total of 42 items: (1) Perceptions regarding the implementation and awareness of AI (10 items), (2) Perceived advantages of AI in healthcare delivery (9 items), (3) Challenges and obstacles to the adoption of AI (11 items), and (4) Culture of patient safety and organizational preparedness (12 items). All items were evaluated using a five-point Likert scale, with ratings ranging from 1 (Strongly disagree) to 5 (Strongly agree). Higher scores reflected more favorable perceptions regarding AI integration and a more robust patient safety culture.
The questionnaire was first created in English and subsequently translated into Arabic using a standardized forward-backward translation process to ensure semantic and conceptual equivalence. A panel of five experts in health informatics, patient safety, and public health conducted a content validity assessment by reviewing the items for clarity, relevance, and cultural appropriateness. Wording adjustments were implemented in accordance with their feedback. A pilot study was conducted with 30 healthcare practitioners from various professional categories in Riyadh to assess the instrument’s clarity and reliability. Pilot responses were utilized solely for reliability testing and were excluded from the final analysis.
The questionnaire’s internal consistency was deemed satisfactory, with a Cronbach’s α of 0.85. The reliability coefficients for the specific domains were acceptable: Perceptions and awareness (α = 0.81), Perceived benefits (α = 0.84), Challenges and barriers (α = 0.78), and Patient safety culture and readiness (α = 0.87). The values demonstrate strong internal consistency, affirming the reliability of the measurement tool within the current context.
Data sources and measurement
Data were collected electronically, ensuring accessibility and ease of participation. The target population was reached through collaboration with hospital administration and professional networks. Emails were obtained through internal mailing lists used by quality departments. Additionally, the survey was shared with professional healthcare groups on platforms such as WhatsApp, including practitioners from governmental, private, and semi-governmental institutions in Riyadh.
To ensure data integrity and compliance with ethical guidelines, participants were informed of the study’s objectives, confidentiality measures, and the policy regarding voluntary participation. Before proceeding with the questionnaire, participants had to provide informed consent, acknowledging their willingness to participate and understanding of the confidentiality policy. Ethical approval was obtained from King Fahad Medical City in Riyadh, Saudi Arabia (25–026 IRB); the current study strictly adhered to all ethical considerations of scientific research.
Bias
To minimize response bias, the survey was designed to be anonymous and widely distributed across various healthcare settings in Riyadh. Convenience sampling may introduce selection bias, but the study sought to mitigate this by targeting a diverse group of healthcare professionals from private, governmental, and Semi-governmental institutions.
Study sample size
The sample size was calculated using Cochran’s formula for cross-sectional survey studies. The calculation was based on a 95% confidence level, a 5% margin of error, and an estimated response proportion of 50%, which was selected to provide the most conservative estimate and maximize variability [15]. Based on these assumptions, the minimum required sample size was 384 participants. To account for potential incomplete or unusable responses, the target sample was increased by approximately 10%, resulting in a target sample of 422 participants. A total of 396 valid responses were included in the final analysis, which exceeded the minimum required sample size [16]. Given that the study was primarily descriptive and did not involve inferential hypothesis testing, the sample size was considered adequate for estimating healthcare practitioners’ perceptions of AI integration and patient safety culture with acceptable precision.
Statistical methods
Data was analyzed utilizing IBM SPSS Statistics version 27.0 (IBM Corp., Armonk, NY, USA). Before analysis, data were examined for completeness and consistency. Responses containing above 20% missing data were omitted from the final dataset. Descriptive statistics were calculated for participants’ demographic attributes and survey responses. Frequencies and percentages were utilized for categorical data, whilst means and standard deviations were employed for continuous or composite variables.
Composite mean scores for each thematic category of the questionnaire—perceptions of AI implementation and awareness, perceived advantages, difficulties and hurdles, and patient safety culture and readiness—were computed by averaging the relevant questions. Elevated ratings signified more favorable attitudes of AI and a more robust patient safety culture. The internal consistency of each domain was evaluated using Cronbach’s α, with values of 0.70 or higher considered satisfactory for dependability.
All analyses were descriptive, and no inferential statistical tests were used, as the primary aim of the study was to elucidate existing attitudes, implementation strategies, and perceived barriers to AI adoption among healthcare practitioners in Riyadh.
Results
Participants
A total of 500 healthcare practitioners were invited to participate in the online survey through institutional email lists, professional groups, and social media platforms. Of these, 452 individuals accessed the survey link, and 424 met the eligibility criteria (i.e., currently employed in healthcare settings within Riyadh). After data cleaning and removal of incomplete submissions with more than 20% missing responses, 396 completed questionnaires were included in the final analysis. This corresponds to an effective response rate of approximately 79%. The main reasons for exclusion were ineligibility (n = 28) and incomplete or partially submitted questionnaires (n = 56).
Descriptive data
The respondents’ demographic characteristics are summarized in Table 1. The majority were female (81%), while males constituted only 19%. Most participants were aged 20–29 (45%) and 30–39 (45%). Most participants were healthcare professionals (clinicians directly involved in patient care) working primarily in private-sector institutions (77%), while the remaining participants worked in governmental (22%) and semi-governmental (2%) sectors.
Table 1.
Respondent’s demographic data
| Variables | Number (n) | Percentage (%) |
|---|---|---|
| Gender | ||
| Female | 321 | 81% |
| Male | 75 | 19% |
| Age | ||
| 20–29 | 180 | 45% |
| 30–39 | 177 | 45% |
| 40–49 | 30 | 8% |
| > 50 | 9 | 2% |
| Professional Role | ||
| Healthcare professionals (clinicians directly involved in patient care) | 347 | 88% |
| Healthcare associate professional | 8 | 2% |
| Healthcare management and support personnel | 12 | 3% |
| Personal care workers in health services | 6 | 2% |
| Others | 23 | 6% |
| Work Setting | ||
| Governmental | 86 | 22% |
| Private | 303 | 77% |
| Semi-Governmental | 7 | 1% |
AI implementation
The respondents’ perceptions of AI implementation in their institutions are detailed in Table 2. The majority of respondents (60% expressed familiarity with AI applications in their work. More than half of the participants agreed or strongly agreed that AI is used in their institution for diagnostic support (55%), treatment planning (51%), patient monitoring (59%), Administrative tasks (56%), and risk management (55%).
Table 2.
Opinions of healthcare providers about AI implementation
| Strongly agree n(%) | Agree n(%) | Neutral n(%) | Disagree n(%) | Strongly disagree n(%) | |
|---|---|---|---|---|---|
| I am familiar with AI applications in healthcare in my work | 42 (11%) | 195 (49%) | 113 (29%) | 29 (7%) | 12 (3%) |
| In my institution, AI is used in diagnostic support | 28 (7%) | 190 (48%) | 108 (27%) | 50 (13%) | 18 (5%) |
| In my institution, AI is used in Treatment planning | 33 (8%) | 171 (43%) | 116 (29%) | 46 (12%) | 25 (6%) |
| In my institution, AI is used in patient monitoring | 43 (11%) | 190 (48%) | 90 (23%) | 49 (12%) | 22 (6%) |
| In my institution, AI is used in administrative tasks | 34 (9%) | 186 (47%) | 109 (28%) | 45 (11%) | 18 (5%) |
| In my institution, AI is used in Risk management | 41 (10%) | 180 (45%) | 108 (27%) | 43 (11%) | 20 (5%) |
Participants’ beliefs about the impact of AI on patient safety
Table 3 presents participants’ beliefs regarding AI’s impact on patient safety. (69%) of participants agreed or strongly agreed that AI improves patient safety. In detail, the majority of participants (70%) believed that AI improves patient monitoring, while 65% of participants felt that AI streamlines administrative processes. 68% thought it reduces diagnostic errors, 69% believed AI enhanced treatment accuracy, 66% believed AI facilitated risk prediction, 67% believed AI enhanced communication and collaboration for patient safety, and 57% believed AI positively impacted patient safety culture. While a substantial portion remained neutral, ranging from 22% to 29%, the “disagree” and “strongly disagree” responses remained consistently low, averaging below 6% per statement.
Table 3.
Respondent’s beliefs of AI’s impact on patient safety
| Variables | Strongly agree n(%) | Agree n(%) | Neutral n(%) | Disagree n(%) | Strongly disagree n(%) |
|---|---|---|---|---|---|
| I believe AI implementation improves patient safety in my organization | 61 (15%) | 215 (54%) | 88 (22%) | 17 (4%) | 9 (2%) |
| I believe AI reduces diagnostic errors | 43 (11%) | 224 (57%) | 99 (25%) | 13 (3%) | 11 (3%) |
| I believe AI enhances treatment accuracy | 39 (10%) | 232 (59%) | 104 (26%) | 11 (3%) | 8 (2%) |
| I believe AI facilitates risk prediction | 36 (9%) | 226 (57%) | 115 (29%) | 9 (2%) | 6 (2%) |
| I believe AI improves patient monitoring | 50 (13%) | 224 (57%) | 97 (24%) | 11 (3%) | 7 (2%) |
| I believe AI streamlines administrative processes. | 42 (11%) | 214 (54%) | 116 (29%) | 10 (3%) | 10 (3%) |
| I believe AI enhances communication and collaboration related to patient safety. | 45 (11%) | 221 (56%) | 98 (25%) | 22 (6%) | 6 (2%) |
| I believe the implementation of AI will positively impact the patient safety culture | 45 (11%) | 221 (56%) | 105 (27%) | 10 (3%) | 10 (3%) |
Challenges and recommendations
Table 4 highlights the challenges and recommendations for using AI. The most frequently reported challenges were concerns about data privacy in AI usage (57%) and a lack of training in utilizing AI for patient safety (55%). More than half of the participants (53%) “agreed” or “strongly agreed” that limited access to AI resources for patient safety applications is also a challenge. Technical difficulties and resistance to change respondents’ rates were 44% and 40%, respectively.
Table 4.
Respondent’s challenges and recommendations to AI
| Variables | Strongly agree n(%) | Agree n(%) | Neutral n(%) | Disagree n(%) | Strongly disagree n(%) |
|---|---|---|---|---|---|
| I am facing a lack of training in using AI for patient safety. | 37 (9%) | 181 (46%) | 121 (31%) | 45 (11%) | 7 (2%) |
| I have limited access to AI resources for patient safety applications | 34 (9%) | 175 (44%) | 131 (33%) | 45 (11%) | 4 (1%) |
| I am facing technical difficulties when using AI tools | 28 (7%) | 147 (37%) | 149 (38%) | 59 (15%) | 9 (2%) |
| I’m concerned about data privacy when using AI for patient safety | 44 (11%) | 184 (46%) | 128 (32%) | 24 (6%) | 6 (2%) |
| There is resistance to change in using AI within my organization | 23 (6%) | 134 (34%) | 182 (46%) | 42 (11%) | 10 (3%) |
Discussion
The present study aimed to examine healthcare practitioners’ perceptions of using AI to enhance patient safety culture in Riyadh, Saudi Arabia. A limited amount of literature was identified, focusing specifically on the perception of AI use for enhancing patient safety culture among healthcare practitioners. The results of this study align with the objectives outlined in Saudi Vision 2030, particularly under the Health Sector Transformation Program [1]. Vision 2030 emphasizes the importance of developing an effective and integrated healthcare system centered on the patient. One of the program’s key pillars focuses on enhancing healthcare by adopting advanced digital technologies, including AI, to improve service efficiency, reduce medical errors, and improve patient outcomes. Identifying healthcare practitioner perceptions will directly support Vision 2030’s goals of leveraging technological innovation [15].
The findings of this study provided a comprehensive understanding of a generally positive outlook toward AI’s role in optimizing patient safety culture, with over half of the participants agreeing that Participants perceived AI as having the potential to improve diagnostic accuracy, reduce medication errors, and streamline clinical workflows. In Table 1, the gender distribution shows that healthcare practitioners are predominantly female, suggesting a gendered trend in perceptions of AI implementation. The age distribution indicates that most respondents are young professionals, which may reflect greater adaptability and openness to AI technologies. Regarding work settings, private-sector employees make up the majority, while government and semi-governmental organization employees account for a smaller share. This distribution may influence AI accessibility, as indicated by the higher proportion of participants from private institutions in our study, which may reflect greater exposure to technological advancements than governmental entities.
Differences across professional roles may also help explain variations in perceptions toward AI implementation and patient safety culture. Healthcare professionals directly involved in patient care represented the majority of the sample and may have demonstrated stronger acceptance of AI because of their greater exposure to clinical decision-support systems, patient monitoring technologies, and workflow optimization tools. In contrast, smaller professional subgroups, such as healthcare support personnel and administrative staff, may interact differently with AI systems depending on their organizational responsibilities and level of technological engagement. These variations suggest that perceptions of AI adoption are influenced not only by institutional readiness but also by the nature of professional responsibilities and daily interaction with healthcare technologies. However, because some professional categories were represented by relatively small numbers, subgroup interpretations should be approached cautiously.
In the present study, 60% of healthcare practitioners reported familiarity with AI applications in their work, indicating moderate awareness and engagement with AI technologies. However, a significant proportion (40%) of respondents expressed either neutrality or unfamiliarity, suggesting the presence of knowledge gaps and highlighting the need for more structured AI training initiatives, as shown in Table 2. Similar trends have been observed in previous studies. Radiologists in Saudi Arabia demonstrated varying levels of AI knowledge and familiarity, with many practitioners recognizing AI’s presence in healthcare but showing limited practical familiarity [17, 18]. In another study, healthcare providers reported a 48.1% familiarity rate with AI applications in their specialty. Although they held a positive perception of AI integration, many expressed uncertainty about their preparedness to implement AI technologies. Additionally, although a substantial proportion feared that AI might replace their roles, it was clarified that AI is seen more as a supportive tool than a replacement for physicians [13]. Dental professionals in Saudi Arabia also demonstrated moderate awareness, with 49.4% reporting familiarity with AI, though practical usage remained low due to insufficient training [19]. On the other hand, community pharmacists demonstrated greater familiarity, with 94.5% reporting awareness of AI technologies; however, formal AI training remained limited [20]. Furthermore, physicians’ familiarity varied; while 76% acknowledged AI’s accuracy and 60% recognized its efficiency, only 25.9% had actually used AI systems in clinical practice, reflecting a similar gap between recognition and application [21].
Regarding treatment planning, as shown in Table 2, approximately 51% of respondents in this study agreed or strongly agreed that AI is used in treatment planning in their work. This reflects a growing yet not fully established integration of AI technologies in clinical decision-making. Previous research supports this trend [3, 12]. AI has been highlighted as instrumental in facilitating personalized treatment plans, particularly through the analysis of complex datasets and the use of predictive algorithms to optimize patient-specific strategies [12]. The potential of AI to support personalized treatment planning was further emphasized, as it can analyze large volumes of patient data to develop treatment plans tailored to medical histories and clinical conditions [17]. Similarly, 56.6% of physicians in another study believed AI should assist in diagnosis and treatment planning. However, they stressed that such systems should operate under physician supervision, with only a minority favoring complete AI autonomy in decision-making [21].
The findings in Table 3 reveal strong consensus among participants on the positive impact of AI on patient safety. 69% agreed or strongly agreed that AI contributes to overall patient safety, highlighting a prevalent perception of AI as a beneficial tool in healthcare. A study conducted in Jeddah, Saudi Arabia, found that 83% of participants believed AI could be helpful in their work [22].
Diagnostic errors are a significant concern in healthcare, contributing to adverse patient outcomes and increased costs. The perception that AI can mitigate these errors highlights a strong belief in its potential to improve diagnostic accuracy and efficiency. The finding that 68% of participants agreed or strongly agreed that AI reduces diagnostic errors is particularly significant. This is supported by findings from a cross-sectional study among healthcare workers in private clinics in Jeddah, Saudi Arabia, which reported that 75% of participants believed AI could help reduce medical errors [22]. In another cross-sectional study, which was conducted to evaluate the knowledge, attitude, and practices regarding AI among registered nurses, nursing faculty, and students in Saudi Arabia, the majority of participants (52.6%) agreed that AI has the potential to improve the accuracy and efficiency of nursing diagnosis and treatment [23].
The findings of this study, presented in Table 4, reveal significant challenges to AI adoption among healthcare practitioners, particularly regarding data privacy and training needs. Concerns about data privacy in AI utilization within healthcare settings have been highlighted in recent studies. Elnaggar et al. reported that 74.6% of participants acknowledged serious privacy issues related to AI use [13], while Chikhaoui et al. further reinforced these concerns, noting that 57% of participants considered data privacy a critical area requiring robust security measures [14]. These findings consistently emphasize the need to address data privacy issues to build trust and ensure the ethical implementation of AI in healthcare.
The lack of sufficient training presents a significant barrier to the responsible integration of AI technologies. Chikhaoui et al. reported that 55% of participants identified insufficient training in AI utilization as a substantial obstacle to ensuring patient safety [14], and Elnaggar et al. noted that a significant majority (76.2%) favored incorporating AI courses into medical curricula [13]. This widespread support underscores the critical need for educational initiatives to equip healthcare practitioners with the skills to integrate AI safely and effectively. Additionally, inadequate and unrepresentative training data can lead to inaccurate AI algorithms, compromising their effectiveness in preventing medical errors [24]. Using sensitive patient data in AI development raises significant privacy and security concerns, necessitating robust safeguards and adherence to ethical guidelines [25, 26]. Addressing these challenges requires collaborative efforts among healthcare providers, researchers, policymakers, and patients to ensure responsible AI development and implementation, with a focus on data security, transparency, and ethical considerations. This approach aims to maximize the benefits of AI while minimizing potential risks.
Overall, the study highlights a generally positive perception of AI’s potential to improve patient safety. However, to achieve successful implementation, it is essential to address key barriers—particularly those related to training and data privacy—which may hinder the effective adoption of AI in healthcare settings.
Conclusions
The study examines healthcare practitioners’ perceptions of using AI to enhance patient safety culture in Riyadh, Saudi Arabia. This study provides a comprehensive understanding of the generally positive outlook toward AI’s role in improving patient safety culture, with over half of participants agreeing that AI was perceived by participants as having the potential to enhance diagnostic accuracy, reduce medication errors, and streamline clinical workflows.
However, substantial challenges remain, most notably those related to data privacy and the need for adequate training in AI adoption. Addressing these challenges through focused training programs, robust data governance policies, and ethical frameworks will enable healthcare systems to adopt AI safely and effectively. Addressing these challenges is therefore crucial to unlocking the full potential of AI to enhance patient safety culture and improve healthcare efficiency—objectives closely aligned with Saudi Arabia’s Vision 2030. Future research should explore strategies to mitigate these challenges, assess the impact of AI adoption on patient safety culture, and investigate the integration of AI across diverse healthcare settings in Saudi Arabia.
Limitations
The current study has several limitations that should be acknowledged to provide a clearer context for the findings. Because the study relied exclusively on self-reported perceptions, findings may reflect subjective beliefs rather than actual AI utilization or measurable patient safety outcomes. Social desirability bias may have influenced respondents to report favorable attitudes toward technology adoption or safety practices. Additionally, perception–reality gaps may exist, in which reported familiarity with or institutional AI use does not fully correspond to operational implementation. These factors limit the ability to infer real-world effectiveness and highlight the need for future studies incorporating objective system metrics, observational methods, or longitudinal designs. A further limitation concerns the gender distribution within the study sample. There was a clear predominance of female respondents, which could have influenced the outcomes, given that gender differences are known to affect beliefs toward technology adoption. The use of a self-administered questionnaire introduces the potential for response bias, including recall bias and social desirability bias. Participants may have either overstated positive perceptions or concerns due to professional or societal expectations. Adopting multiple data collection methods would help minimize such bias in future research. Lastly, as a cross-sectional study, the research captures data at a single point in time, which limits the ability to determine causal relationships between variables. This design does not account for temporal changes or long-term effects; therefore, findings should be interpreted as associations rather than causal relationships. Finally, Professional role categories were grouped into broad categories, limiting the ability to conduct more detailed, stratified analyses across specific clinical professions. This restricts the nuanced interpretation of how perceptions may differ among occupational groups. Future studies should collect more granular professional classifications to allow deeper subgroup analyses. Also, Future research should recruit more proportionally balanced professional subgroups to allow more robust comparative analyses across healthcare disciplines.
Recommendations
The following recommendations are proposed to fully realize AI’s potential to enhance patient safety in the Saudi Arabian healthcare system. First, comprehensive AI training programs should be developed and implemented specifically for healthcare practitioners to bridge the knowledge and skill gaps identified in this study. Second, data privacy and security frameworks should be strengthened to address concerns about the confidentiality and protection of patient information and ensure compliance with ethical standards. Third, address resistance to change by fostering a culture that embraces technological innovation, including involving healthcare professionals in the AI implementation process and promoting awareness of AI’s benefits for patient safety. Fourth, expand future research efforts by including larger, more diverse samples across various regions of Saudi Arabia and employing a combination of quantitative and qualitative data collection methods. This will enhance the robustness and generalizability of findings while providing deeper insights into the perceived relationship between AI integration and patient safety. Aligning these initiatives with the strategic objectives outlined in Saudi Vision 2030 will support ongoing healthcare modernization efforts and contribute to improved patient care and safety across the Kingdom.
Supplementary Information
Below is the link to the electronic supplementary material.
Abbreviations
- AI
Artificial intelligence
- HCPs
Healthcare practitioners
- IRB
Institutional review board
- MOH
Ministry of health
- SD
Standard deviation
- SPSS
Statistical package for the social sciences
Author contributions
Amal Abuabat (AA) and Aeshah Almutairi (AAI) contributed to study conceptualization, data collection, and manuscript drafting. Noura Gazali (NG) contributed to data analysis and interpretation. Dr. Salah Alshagrawi (SA) supervised the study design, methodology, and manuscript development, and provided critical revisions. All authors reviewed and approved the final manuscript.
Funding
No funding was provided to the authors.
Data availability
The data supporting the findings of this study are available from the corresponding author, SA, upon request. The data is not publicly available because it contains information that could compromise the privacy of research participants.
Declarations
Ethics approval and consent to participate
Ethics approval and consent to participate: Ethical approval for this study was obtained from the Institutional Review Board (IRB) of King Fahad Medical City, Riyadh, Saudi Arabia (IRB Reference No. 25–026). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki and all applicable local regulations governing research involving human participants. Electronic informed consent was obtained from all participants prior to participation in the study.
Consent for publication
We confirm our approval for publication.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Mani ZA, Goniewicz K. Transforming healthcare in Saudi Arabia: a comprehensive evaluation of Vision 2030’s impact. Sustainability. 2024;16(8):3277. 10.3390/su16083277. [DOI] [Google Scholar]
- 2.Al-luhaym AN, Alshagrawi S. Factors associated with access to the saudi primary healthcare in light of vision 2030. Sudan J Med Sci. 2023;18(3):391–401. [Google Scholar]
- 3.Saeed A, Bin Saeed A, AlAhmri FA. Saudi Arabia health systems: challenging and future transformations with artificial intelligence. Cureus. 2023;15(4):e37826. 10.7759/cureus.37826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Alamri HM, Alshagrawi S. Factors influencing telehealth adoption in managing healthcare in Saudi Arabia: a systematic review. J Multidiscip Healthc. 2024 Dec;31:5225–35. [DOI] [PMC free article] [PubMed]
- 5.Al Kuwaiti A, Nazer K, Al-Reedy A, Al-Shehri S, Al-Muhanna A, Subbarayalu AV, et al. A review of the role of artificial intelligence in healthcare. J Pers Med. 2023;13(6):951. 10.3390/jpm13060951. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Alqurashi MA, Alshagrawi S. Assessing the impact of artificial intelligence applications on diagnostic accuracy in Saudi Arabian healthcare: a systematic review. Open Public Health J. 2025;18(1):225–234.
- 7.Bajwa J, Munir U, Nori A, Williams B. Artificial intelligence in healthcare: transforming the practice of medicine. Future Healthc J. 2021;8(2):e188–94. 10.7861/fhj.2021-0095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Mehta V. Artificial intelligence in medicine: revolutionizing healthcare for improved patient outcomes. J Med Res Innov. 2023;7(2):e000292. [Google Scholar]
- 9.Alsulami A, A’aqoulah A, Almutairi N. Patient safety culture awareness among healthcare providers in a tertiary hospital in Riyadh, Saudi Arabia. Front Public Health. 2022;10:953393. 10.3389/fpubh.2022.953393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Alahmari MH, Alshagrawi S. Examining biomedical waste management knowledge and training level among health care professionals in Saudi Arabia. Am J Infect Control. 2024;52(9):1012–9. [DOI] [PubMed] [Google Scholar]
- 11.Mistri IU, Badge A, Shahu S. Enhancing patient safety culture in hospitals. Cureus. 2023;15(12):e51159. 10.7759/cureus.51159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Alowais SA, Alghamdi SS, Alsuhebany N, Alqahtani T, Alshaya AI, Almohareb SN, et al. Revolutionizing healthcare: the role of artificial intelligence in clinical practice. BMC Med Educ. 2023;23:689. 10.1186/s12909-023-04698-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Elnaggar M, Alharbi ZA, Alanazi AM, Alsaiari SO, Alhemaidani AM, Alanazi SF, et al. Assessment of the perception and worries of Saudi healthcare providers about the application of artificial intelligence in Saudi health facilities. Cureus. 2023;15(8):e42858. 10.7759/cureus.42858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Chikhaoui E, Alajmi A, Larabi-Marie-Sainte S. Artificial intelligence applications in healthcare sector: Ethical and legal challenges. Emerg Sci J. 2022;6(4):05. 10.28991/ESJ-2022-06-04-05. [DOI] [Google Scholar]
- 15.Sample size calculator by Raosoft, Inc. http://www.raosoft.com/samplesize.html. Accessed 20 Apr 2025.
- 16.Health sector transformation program. https://www.vision2030.gov.sa/en/vision-2030/vrp/health-sector-transformation-program/. Accessed 9 Oct 2023.
- 17.Alghamdi SA, Alashban Y. Knowledge, attitudes and practices towards artificial intelligence (AI) among radiologists in Saudi Arabia. J Radiation Res Appl Sci. 2023;16(2):100569. 10.1016/j.jrras.2023.100569. [DOI] [Google Scholar]
- 18.Alqahtani B, Alshagrawi S. The Relationship between Leadership Style, Employment Engagement, and Turnover Intentions in Healthcare Organizations in Saudi Arabia. Majmaah J Health Sci. 2025;13(1):116. [Google Scholar]
- 19.Aboalshamat KT. Perception and utilization of artificial intelligence (AI) among dental professionals in Saudi Arabia. Open Dentistry J. 2022;16:e2208110. 10.2174/18742106-v16-e2208110. [DOI] [Google Scholar]
- 20.Syed W, Al-Rawi MBA. Community pharmacists’ awareness, perceptions, and opinions of artificial intelligence: A cross-sectional study in Riyadh. Saudi Arabia Technol Health Care. 2024;32(1):481–93. 10.3233/THC-230784. [DOI] [PubMed] [Google Scholar]
- 21.Alkhatieb M, Subke AA. Artificial intelligence in healthcare: A study of physician attitudes and perceptions in Jeddah, Saudi Arabia. Cureus. 2024;16(3):e57256. 10.7759/cureus.57256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Serbaya SH, Khan AA, Surbaya SH, Alzahrani SM. Knowledge, attitude and practice toward artificial intelligence among healthcare workers in private polyclinics in Jeddah, Saudi Arabia. Adv Med Educ Pract. 2024;15:269–80. 10.2147/AMEP.S448422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mariano MEM, Shahin MAH, Ancheta SJ, Kunjan MV, Al Dossary NM, Al Ojaimi SF, Al Qudah SA, Al Harbi HF. Exploring artificial intelligence knowledge, attitudes, and practices among nurses, faculty, and students in Saudi Arabia: A cross-sectional analysis. Social Sci Humanit Open. 2025;11:101384. 10.1016/j.ssaho.2025.101384. [DOI] [Google Scholar]
- 24.Chen JH, Asch SM. Machine learning and prediction in medicine—Beyond the peak of inflated expectations. N Engl J Med. 2017;376(26):2507–9. 10.1056/NEJMp1702071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Topol EJ. High-performance medicine: The convergence of human and artificial intelligence. Nat Med. 2019;25(1):44–56. 10.1038/s41591-018-0300-7. [DOI] [PubMed] [Google Scholar]
- 26.Albaqami NM, Alshagrawi S. Patient satisfaction with primary health care services in Riyadh, Saudi Arabia. Int J Gen Med. 2025 Dec 31:835–45. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data supporting the findings of this study are available from the corresponding author, SA, upon request. The data is not publicly available because it contains information that could compromise the privacy of research participants.