Adaptive AI systems embedded in clinical decision-making workflows operationalize hidden value tradeoffs at scale, generating distinct liability pressures and constraining clinicians’ ability to weigh and choose among competing value priorities. To redirect attention from technical model design to procurement and implementation accountability, we propose a collaborative framework: federal mandates for standardized transparency via Model Cards, paired with internal, multidisciplinary institutional reviews to deliberate and document value-laden configurations.
Patil et al. (Mon,) studied this question.