Keywords: Artificial intelligence, Healthcare, Antibiotic prescribing, Perceptions, Curriculum

Rajput NR, Eatsham MU, Niama H3, Eman M, Bashir H, Sajjad M. Awareness, Knowledge, Attitudes, and Assumptions about AI in Clinical Practice among Final-Year Medical Students and Internees in Peripheral Hospitals of Sialkot, Pakistan. J Pak Soc Intern Med. 2026;7(3):

Corresponding Author: Dr. Muhammad Umar Eatsham Email: Pharmacist.pioneer@gmail.com

DOI: https://doi.org/10.70302/jpsim.v7i3.2637

Artificial intelligence (AI) is a technology that finds more and more applications in the healthcare field, such as diagnostic imaging, risk-based prediction, and personalized treatment. It enhances study of medical images, risk assessment of patients and personal therapy.1 Antimicrobial stewardship (AMS) is another area supported by AI as a decision support tool, faster diagnostics, and improved treatment. Machine learning could be used to analyze bulk data to identify trends and anticipate antimicrobial resistance (AMR) and patient response to treatments to promote judicious prescribing, improve patient outcomes, and reduce spread of antimicrobial resistance.2 The global health risk that has led to AMR, treatment failures, severe infections and higher mortality; and has also created economic and healthcare burdens is antibiotic misuse.3

The digital health tools are adopted by medical students and young doctors in the beginning of their careers. They need to be aware, perceive and willing to implement AI in clinical decisions in the development of the curriculum and safe AI use. Although increasing evidence is available concerning AI in healthcare, there is still a dearth of research information about the perceptions of young physicians, especially with regard to antibiotic prescribing.4

Generative AI can revolutionize medical education, including admissions, personalized learning, assessment, and research, and there should be a better approach to AI literacy in medical education.5 AI has the potential to support undergraduate education with personalized feedback, adaptable simulations, and virtual environments, providing real-world assessment, increased efficiency, less expense, and safer clinical training6. The obstacles to further implementation are technical infrastructure, privacy issues around data, lack of representation in curricula, and the lack of integration in current learning systems.6

The purpose of this study is to evaluate the level of awareness, attitudes and perceptions of final-year medical students and house officers about AI in healthcare, and its application in the selection of antibiotics.

The study protocol was approved by Islam Medical and Dental College, Sialkot. Participation was voluntary, with informed consent obtained, and responses were anonymized to ensure confidentiality.

The data was analyzed by descriptive statistics, summarizing categorical variables in terms of frequencies and proportions. There were no inferential tests, because the study was to measure awareness and perceptions. Data has been prepared and tabulated in Microsoft Excel reporting results as the number of responses; the number of missing values was not included to percentages.

Out of 400 respondents, the majority of them were 24-26 years old (28.5%), followed by 24-26 (26.5) and 26-26 (25.5); the least aged 27-27 (19.5).

The demography of 400 respondents was as follows: 189 (47.3 percent) of the respondents were male, the number of female respondents was 211 (52.7 percent). Such distribution is close to evenness in general, which increases the external validity of the sample in both genders.

The present study yields the new information about the attitudes of the final-year medical students and internees/house officers to the application of artificial intelligence (AI) in healthcare facilities. The most important result is the overall readiness to embrace AI as an assistive tool in prescribing and clinical decision-making, with the majority of the respondents seeing its future potential, although they still demanded the established significance of physician supervision. These findings suggest the middle-ground standpoint: artificial intelligence is not meant to substitute clinical judgment, but, rather, supplement it, in accordance with the dynamic role of digital technology in clinical practice.

The one interesting finding of the present research study is that there is no formal education on artificial intelligence (AI) and digital health in contemporary medical education. Although the study participants mostly did not have any organized teaching experiences in the field, most of them were interested in AI-related workshops and welcomed the introduction of AI and digital health issues into the curriculum and similar kind of scenario was observed in a published study.7 This highlights a clear gap between the technological demands of modern healthcare and the preparedness of future physicians, underscoring the need for curriculum reform. Similar calls for integration of AI competencies into medical training have been noted in recent studies, reflecting a growing recognition of digital health literacy as a core component of medical professionalism.8

A large number of participants were aware of use of AI in healthcare settings and numbers goes up to 80% and the similar kind of results were identified by a study published in China 9. Despite of being aware by the use of AI in healthcare, a large proportion of respondents e.g more than 80% did not receive any formal training during their formal medical education and remaining 20% received AI training on their own. Similar results are observed by a study published in Canada although an overall excitement currently surrounds the medical student population, recent surveys indicate that most of these students have not, in fact, been taught anything about artificial intelligence. As of 2022, just 15 per cent of Canadian medical students had so far encountered AI-related course material.7 The necessity to integrate the skills of artificial intelligence into medical education has also been identified in recent studies, as the digital health literacy has come to be the critical component of the medical professionalism.10,11,12 This study revealed that the participants despite having no formal training during their medical education, privately explored the AI and it’s possible in their medical career. Similar kind of scenario was observed in a published study in America that students took part in independent studies of artificial intelligence, relieving them of the need to be supervised by faculty.8 The information obtained in the course of the current study indicates that most of the respondents are confident that the implementation of artificial intelligence can be used to improve antibiotic prescription habits and lead to a decrease in antimicrobial resistance and several studies support this claim as well.13,14 The findings of this study suggest that over 70 % of the participants support the inclusion of artificial intelligence in the curriculum of medical professionals or the delivery of specific training on the topic of artificial intelligence as is described in a study carried out in the United States.8

The current study has great academic and clinical relevance. Artificial intelligence (AI) and the associated digital health modules should be included in undergraduate and postgraduate curricula in medical education to enhance responsible use and equip future physicians to critically assess and implement AI tools. At the same time, the results of the study show that young doctors are in a good position to embrace AI and they need protective mechanisms, including transparency, systematic monitoring, and extensive education.

There are a few limitations to the current study. First, the cross-sectional study design does not allow drawing of causal conclusions. Second, the data rely on self-reported perceptions, which can be idiosyncratic personal exposure to technology or current norms of education at the institutions of the respondents. Third, the findings are narrowed to a certain geographic and institutional setting; therefore, they cannot be generalized. Future studies should incorporate studies that are larger and multi-center, longitudinal in nature, thus tracing how attitudinal dimensions evolve as a result of actual exposure to AI in clinical practice.

The researchers find that undergraduate students in their final year and house officers are aware of AI potential to enhance healthcare, but it is essential that physicians have supervision over AI. Despite the lack of formal training, the participants are interested in AI education and want to see it implemented in the curricula. Significant ethical and clinical issues, such as data privacy, accuracy and empathy, should be addressed prior to safe implementation. Overall, the findings suggest that the future physicians will be enthusiastic and cautious, and with formal training, effective protective strategies, and further exploration, the application of AI in the clinical setting can be pursued safely.

Ethical Approval: The IRB/EC approved this study via letter no. 224/DME/KMC Dated April 13, 2023.

Conflict of Interest: None

Funding Source: None