Abstract:

Introduction:
Rapid development of artificial intelligence (AI) is significantly impacting medical education (ME). AI systems, like Chat Generative Pre-trained Transformer (ChatGPT) from OpenAI, mimic the human brain’s ability to process vast amounts of complex data. Since 2023, ChatGPT has gained popularity worldwide in ME for simulating personal tutoring. It produces responses that are systematic, coherent, and informative, responding in a human-like manner to users’ language inputs. Anatomy learning often overwhelms medical students (MS), particularly neuroanatomy, due to its complexity. In this context, we aim to assess ChatGPT’s effectiveness in anatomy education (AE), particularly in neuroanatomy teaching (NAT).

Methodology:
We reviewed literature from 2011–February 2024, using scientific electronic databases and relevant keywords to gather pertinent data.

Results:
ChatGPT has been tested in several anatomical areas, including the head, neck, back, pelvis, thorax, and musculoskeletal system. ChatGPT-4(latest version) demonstrated its ability to organise information on anatomical features and draw connections between anatomy, physiology, and clinical concepts. It shows promise in revolutionising AE by reducing the time educators and students spend on teaching, learning tasks, and exam preparation.

ChatGPT-4 has various features that can aid NAT:
- Detailed explanations on complex human anatomy, enhancing engagement and understanding
- Personalised responses to students’ questions for tailored learning
- Interactive “virtual doctor patient consultations” for practicing communication and diagnostic skills
- Simulated scenarios (SS) for problem-based learning, reinforcing complex neuroanatomy for real-life situations
- Custom study guides and summaries, focusing on intricated neuroanatomy structures for easier recall

Future NAT will benefit from including neuroanatomy, neurology case scenarios, and neuroimaging interpretation aiming to enhance three-dimensional neuroscience learning to overcome the associated cognitive overload. ChatGPT-4 is well-suited for these tasks, offering time and cost savings.

Advantages include personalised learning support, generation of concise summaries and exam questions variations, and round-the-clock chatbot availability, catering to different learning schedules. Disadvantages involve inaccuracies (72.5% accuracy rate), algorithm bias, and challenges such as translating complex anatomy content into different languages and creating inaccurate anatomical images. ChatGPT’s responses cannot be deemed as evidence-based due to lack of citations, and its latest knowledge update being in September 2021. Limitations-Continuous updates and manual retraining to align with research are necessary. Additionally, the ethical implications of AI in ME concerning accuracy and validity, as well as potential for inequality due to differences in digital access, need consideration.

Conclusion:
ChatGPT-4 could be considered a valuable interactive anatomy tool for MS to enhance individual learning abilities by engaging and motivating. It is a useful tool for exam preparation, helping to alleviate the stress of memorising extensive anatomical details. AI could be seen as a complementary tool in AE, supporting both educators’ and students’ learning. Further research into creating standardised guidelines for ChatGPT-4 use in AE and curriculum designing is needed. Policies should address potential learning inequalities.

Keywords:

ChatGPT, Anatomy, Artificial Intelligence, Anatomy Education, Medical Students, Medical Education.