ChatGPT in 21st century medicine: advances, ethical challenges and inherent limitations

Abstract

The third decade of the 21st century highlights technological advancements in the field of Artificial Intelligence (AI), notably with OpenAI’s ChatGPT, a tool based on Natural Language Processing (NLP) on an unprecedented scale. ChatGPT finds applications across various sectors, including medicine, where it can assist in discussing clinical cases and suggesting differential diagnoses, as well as providing translations, definitions, and personalized guidance, among other potential applications. However, there are limitations, including internet data bias and the inability to fully comprehend content. The tool may also generate incorrect responses and be sensitive to question formulation, in addition to lacking knowledge updates beyond 2021. Importantly, ethical and security considerations are crucial, as AI trainers can review conversations, necessitating the assurance of medical data confidentiality. While ChatGPT holds revolutionary potential in medicine, its implementation requires human supervision, ethical considerations, and attention to its inherent limitations.