Artificial intelligence (AI) is poised to revolutionize medical education, offering innovative tools and resources for students. Given that approximately 44% of children utilize AI for schoolwork/homework [1], this review investigates AI's potential applications within medical education and evaluates its benefits and drawbacks.

AI, encompassing technologies like natural language processing, machine learning, computer vision, and generative AI [2], can transform medical education. AI's integration into online question banks, such as Passmedicine, personalizes learning through spaced repetition and personalized feedback, enhancing learning outcomes. Spaced repetition, informed by Ebbinghaus' Forgetting Curve [3], and personalized feedback [4], are further optimized by AI.

AI also enhances clinical skills training. Large language models (LLMs), like ChatGPT, generate realistic patient scenarios for history-taking practice. The voice conversation feature in LLMs enables real-time interaction, simulating patient consultations [5].

Advanced AI techniques, particularly Natural Language Learning Models (NLLMs), offer further development avenues [6]. NLLMs, like virtual chatbot assistants, personalize learning by tailoring answers to individual comprehension levels. While a study suggests NLLMs may not directly enhance information retention, they can increase student motivation when engaging with challenging topics [7].

AI-driven transcription tools improve accessibility for students with learning differences, such as dyslexia [8], promoting a more equitable learning environment. However, the risk of over-reliance on AI tools is a concern, and research is needed to ensure accuracy in medical terminology and address the development of essential practical skills.

Machine learning (ML) and deep learning (DL) AI models offer unique capabilities for analysing data and providing skills and performance feedback. ML excels at pattern recognition, while DL processes images and recognizes complex patterns, as seen in surgical technique optimization. Studies show ML and DL models can be used to classify skill levels and provide personalized feedback, and VR simulations with AI have shown significant performance improvements in surgical training [9, 10, 11]. However, these applications are primarily focused on postgraduate clinicians, highlighting the need for adaptations for undergraduate medical students, such as AI simulations for basic procedures.

Ethical considerations are crucial. AI algorithms can be biased, potentially leading to healthcare discrimination. Data and privacy concerns must be addressed to ensure patient confidentiality.

In conclusion, while AI supplementation in medical education holds promise, further research is needed to address ethical challenges, ensure accuracy, and optimize its application for medical students, especially in acquiring clinical skills.

Keywords: Artificial intelligence, Medical-education, medical students, technology enhanced learning, large language learning models, machine learning, deep learning.