QuestionQ96

VerifiedDomain IV — Understanding how to govern AI deployment and use

A US hospital intends to develop an AI system that reviews available patient data to propose an initial diagnosis to licensed physicians. The hospital will adopt a policy requiring physicians to consider the AI’s proposal, while conducting their own physical examinations before making a final diagnosis.

An important ethical concern with this plan is:

Explanation

Training a diagnostic AI on a representative dataset is essential to reduce biased performance across different patient populations. Underrepresentation can lead to less accurate or inequitable diagnostic recommendations for particular groups, creating a significant clinical ethics and fairness risk.

Community Discussion

No comments yet. Be the first to start the discussion!