Diagnosis – AI or Doctor? (DiAD) Investigating the Ethical and Practical Implications of Patient use of Chatbot AIs Compared with Human Doctors in Internal Medicine

Released in November 2022, ChatGPT gained over 100 million users in just two months to become the fastest growing application of all time. This Natural Language Processing Artificial Intelligence is capable of producing human-like responses on an incredibly large range of topics. It is inevitable that ordinary people will use it for medical diagnostic purposes, much as they do any other search engine. Although it, and other NLP tools like it, are comparable in diagnostic accuracy to human doctors, their use as substitutes to, or surrogates for, human doctors raises numerous ethical challenges. Not least of which is the impact this will have on vital doctor-patient relationships in under-served communities.

Current research focuses on quantitative research which does not take into consideration real-world patient experiences. The proposed project will provide an in-depth analysis of real-world patient experiences of using NLP tools in comparison to human doctors.

Following a scoping review, the project will use a novel qualitative research methodology. Real-world patients will be interviewed before a doctor’s consultation. They will then immediately engage in a similar mock consultation with an NLP tool. Data will be subjected to thematic analysis and recommendations produced for NLP tool developers and Swiss policy makers.