Bristol researchers say medicine already knows how to handle black boxes and AI could learn from it.

Researchers at the University of Bristol propose a framework for making medical AI systems safer by borrowing from pharmaceutical drug approval processes. The framework, called Learning Ensemble, defines three areas to check: system limits, fairness across patient groups, and clinical fit.

Why it matters: It suggests a concrete regulatory approach for evaluating AI in medicine based on an established approval model.

Written by AI from The Decoder's published text.