Extrinsic trust as a contractual framework for accountable AI in healthcare: A viewpoint

Artificial intelligence promises efficiency and equity in healthcare. However, adoption remains fragmented due to weak foundations of trust. This Perspective highlights the gap between intrinsic trust based on interpretability, and extrinsic trust based on functional validation. A contractual framework between the AI system and the user is proposed, consisting of three promises: reliability, scope & equity, and shift & uncertainty. Illustrated through a vignette, we argue that health systems require structured evidence, governance frameworks, and data infrastructure to embed these promises, ensuring accountable, safe, and equitable AI deployment.