Research suggests that large language models (LLMs) such as ChatGPT can address a critical gap in teaching ethics to medical students.
A paper published in the journal British Medica Education argues that since there is a lack of educational resources in medical ethics education, including AI ethics in medical curricula can be a “second-best” tool.
The researchers from the Graduate School of Humanities and Social Sciences at the Hiroshima University in Higashi-Hiroshima, Japan [emphasise](https://bmcmededuc.biomedcentral.com/articles/10.1186/s12909-025-06801-y) the critical importance of instilling virtue in medical ethics education.
Addressing the limitations of current LLM models which use large amounts of text data and algorithms to learn patterns, the study suggests the use of carefully constructed prompts to focus on virtue cultivation. If not on their own, LLMs can be used in a hybrid earning model as educational resources on ethics are inadequate.
While admitting that LLMs are not a replacement for teachers, the study uses references from some recent studies that suggest that ChatGPT can exhibit a nuanced understanding of empathy, potentially exceeding human capabilities in recognising emotional subtleties within scenarios.
This indicates that LLMs could offer more empathetic insights in educational settings where students engage with morally complex scenarios.
It also notes the shifting in attitudes of students who do not enter the medical education system (as elsewhere) to uncritically adopt what their teachers tell them but seek to have a ‘reformist ethos’. Thus, virtue ethics are inherently dynamic and adapt over time through critique, dialogue, and attention to changing societal norms.
The authors make it clear that LLMs may be suitable for the classroom, but they are still not to be deployed in actual medical settings. The critical thinking required for a decision in the clinic demands diverse moral perspectives, and LLMs need more training in this regard. Training data for this purpose could come from the education settings, which the essay sees as deficient.
Tsutomu Sawai, one of the authors, said that while LLMs were ready, it was still too early to use them as the ‘definitive sources’ for ethics education.