In recent decades, artificial intelligence (AI) has become a defining feature of the fourth industrial revolution, gaining significant global attention ( Ahmed et al., 2022 ). AI refers to software systems designed to process data and make independent decisions or support human decision-making, mimicking human cognition in learning and problem-solving ( Swed et al., 2022 ). A key subset of AI, Machine Learning (ML), focuses on developing algorithms that improve accuracy through data analysis and pattern recognition ( Martín Noguerol et al., 2019 ; Lee et al., 2019 ). Within ML, Deep Learning (DL) has gained substantial prominence, especially in the medical field. By utilizing artificial neural networks, DL processes and analyzes vast datasets, proving particularly effective for tasks like image processing, diagnostics, and even robotic surgery ( Egger et al., 2022 ).

AI has become a cornerstone of daily life, powering applications such as behavioral algorithms in phone conversations, personalized shopping recommendations on platforms like Amazon and eBay, self-driving cars, and virtual assistants like Google Alexa, Siri, and ChatGPT ( Malodia et al., 2024 ). As AI systems evolve to become more user-friendly and reliable, they have the potential to significantly enhance quality of life ( Ahmad et al., 2023 ). Beyond consumer applications, AI has found widespread use across sectors such as computer science, industrial manufacturing, cybersecurity, finance, and law ( Doumat et al., 2022 ).

In healthcare, AI’s influence has rapidly expanded, revolutionizing technologies such as radiography, medical diagnostics, remote patient monitoring, and drug development ( Murdoch & Detsky, 2013 ). For instance, in oncology, AI assists with cancer diagnosis and grading ( Londhe & Bhasin, 2019 ), while in gastroenterology, it aids in detecting pathological lesions via endoscopic imaging ( Alagappan et al., 2018 ). Additionally, AI is being incorporated into medical education through case-based online learning platforms ( Khumrina et al., 2017 ). Radiology, in particular, has undergone a transformation, as machine-based technologies have become integral to the field ( Paré et al., 2023 ).

Despite its many benefits, concerns persist regarding AI’s impact on the medical profession. Some studies suggest that AI could dampen enthusiasm among medical students, especially those in radiology, due to fears of job displacement. In the United States, 57% of medical students reported that online publications were their primary source of AI knowledge and expressed concerns that AI might replace radiologists, leading to negative perceptions of the technology ( Park, Yi, & Siegel, 2021 ). Similarly, a study in Lebanon found generally low AI knowledge among medical students, though those who received university-level training exhibited significantly higher levels of understanding ( Doumat et al., 2022 ).

Research on the knowledge, attitudes, and practices (KAP) of AI among medical students in Africa is still limited. A study in North Africa revealed low knowledge levels about AI and DL, a moderate perception of AI, and a strong desire for its inclusion in the medical curriculum ( Allam et al., 2023 ).

In Zambia, there is a lack of data on the knowledge, attitudes, and practices (KAP) of medical students regarding AI. A recent study among pharmacy students revealed limited knowledge, negative attitudes, and poor practices related to AI ( Mudenda et al., 2024 ). As such, this study assessed the KAP of AI among medical students at the University of Zambia.

To the best of our knowledge, this is the first study to assess medical students’ knowledge, attitudes, and practices regarding AI in Zambia. Our findings indicate that most medical students at the University of Zambia are aware of AI and demonstrate good knowledge, positive attitudes, and appropriate practices. However, notable gaps exist in their understanding of AI applications in the medical field, highlighting the need for greater integration of AI education in the medical curriculum.

Of the 335 participants, 57.6% (n = 193) were male, while 42.4% (n = 142) were female. The majority (52.5%, n = 176) were between 18 and 22 years old, with fewer participants in older age groups. Most students (77.0%, n = 258) were government-sponsored, and the largest proportion (40.3%, n = 135) were in their third year of study. Additionally, 93.4% (n = 313) of respondents were unmarried. These demographic trends align with findings from similar studies worldwide. For instance, a study conducted in Jordan reported that most participants were female (63.8%), aged between 21 and 22 years, and predominantly in their fifth year of study ( Al-Qerem et al., 2023 ). Likewise, research in Nepal found that 57.9% of participants were female, with the highest representation (20.8%) among fifth-year students ( Jha et al., 2022 ). These similarities suggest that medical student demographics across different regions share common characteristics, particularly in terms of gender distribution and academic progression.

This study found that the majority of participants (94.3%, n = 316) had at least basic knowledge of AI, while 5.7% (n = 19) had no prior exposure. However, only 26.6% were aware of AI applications in medicine, and just 17.9% had knowledge of machine learning (ML) and deep learning (DL), highlighting significant gaps in the medical curriculum. Our findings align with studies from other regions. In Syria, 70% of participants had basic AI knowledge, 23.7% were aware of its medical applications, and 34.7% had adequate understanding of ML and DL ( Swed et al., 2022 ). Similarly, a study in Pakistan reported that 71.3% of students had fundamental AI knowledge, 23.2% understood its applications, and 35.3% were familiar with ML and DL ( Ahmed et al., 2022 ). Additionally, research in Lebanon also found a high level of AI awareness among medical students ( Kharroubi et al., 2024 ).

In contrast, findings from the MENA region revealed significant knowledge gaps, with 87.1% of students demonstrating poor AI knowledge and 83.7% having limited familiarity with ML ( Allam et al., 2023 ). A separate study in Pakistan reported that 61.7% of medical students had no prior AI knowledge but expressed enthusiasm about its integration into their training ( Abid et al., 2019 ). Therefore, these findings underscore the need for structured AI education within medical curricula to bridge existing knowledge gaps and better prepare future healthcare professionals for AI-driven advancements in medicine.

The study found that 77.0% (n = 258) of participants had a positive attitude toward AI, while 23.0% (n = 77) held negative views. More than half (54.6%) agreed that AI is essential in medicine, 48.7% supported its inclusion in the medical curriculum, and 32.5% believed it could eventually replace humans in certain roles. These findings are consistent with studies conducted in other regions. In Pakistan, 47.4% of participants recognized AI’s importance in medicine, and 89% were optimistic about its implementation ( Ahmed et al., 2022 ). A study in the UK also reported that medical students acknowledged AI’s significant role in the field of medicine ( Sit et al., 2020 ), while in Lebanon, 95.6% of students believed AI would have a substantial impact on healthcare ( Doumat et al., 2022 ). Additionally, another study in Pakistan found that 60.8% of students considered AI an effective learning tool, and 58.4% found it credible in medical education ( Sami et al., 2025 ). Similarly, Abid et al. (2019) reported that most medical students in Pakistan had positive attitudes toward AI.

Negative attitudes toward AI in this study appeared to stem primarily from concerns about AI replacing human jobs, a misconception likely influenced by the lack of formal AI education. This aligns with findings from a study conducted among pharmacy students in Zambia, which highlighted similar fears and knowledge gaps ( Mudenda et al., 2024 ). These results underscore the need to integrate AI education into medical training to dispel misconceptions and equip future healthcare professionals with the necessary skills to harness AI’s potential in clinical practice.

In this study, 93.7% of participants demonstrated good AI-related practices, while 6.3% exhibited poor practices. Notably, 90.1% had used AI technologies across various fields, 93.1% acknowledged that AI simplified tasks, and 61.5% expressed interest in working with AI in the future. These findings indicate a higher level of AI engagement compared to studies conducted in other regions. For instance, in Syria, only 78.6% of participants expressed willingness to adopt AI in the future ( Swed et al., 2022 ). Similarly, a study in Pakistan found that while 62.3% of participants were open to AI adoption, only 11.3% had practical experience using AI ( Ahmed et al., 2022 ). The comparatively higher AI exposure and positive attitudes observed in this study highlight a growing awareness and readiness to integrate AI into medical practice, emphasizing the need for structured AI education within medical curricula.

The findings of this study, along with similar research, highlight the growing necessity of integrating AI into medical education, as reported in previous studies ( Oluwadiya et al., 2023 ; Kolachalama and Garg, 2018 ; Chan and Zary, 2019 ; Pinto dos Santos et al., 2019 ). AI has the potential to enhance clinical decision-making, medical diagnostics, and personalized treatment approaches, making it an essential component of modern medical training. Given the increasing role of AI in healthcare, there is a pressing need for its systematic incorporation into medical curricula worldwide. Implementing AI education at an early stage will equip future healthcare professionals with the necessary skills to leverage AI-driven technologies effectively, ultimately improving patient care and healthcare outcomes on a global scale.

We acknowledge some limitations in this study. First, as it was conducted at a single university among medical students, the findings may not be generalizable to the student population at the University of Zambia or to other institutions. Second, the cross-sectional design limits the ability to establish causality or gain an in-depth understanding of the observed trends. Future research should incorporate a mixed-methods approach and longitudinal studies to provide more comprehensive insights. Nevertheless, this study offers valuable baseline information on AI awareness among medical students at the University of Zambia.

This study found that undergraduate medical students generally demonstrated good knowledge, positive attitudes, and appropriate practices regarding AI. However, gaps were identified in their understanding of AI applications in medicine, likely due to the absence of AI in Zambia’s formal medical curriculum. To keep pace with global technological advancements, integrating AI into medical education is essential to better prepare future healthcare professionals.