Automation Bias
Human-AI collaboration is presented as condition-dependent rather than uniformly beneficial or harmful. Clinicians need explicit rules for relying on, verifying, overriding, and remaining independently capable when using clinical AI. Train…
1 sources - 4 claims
Human-AI collaboration is presented as condition-dependent rather than uniformly beneficial or harmful. Clinicians need explicit rules for relying on, verifying, overriding, and remaining independently capable when using clinical AI. Trainees may accept AI outputs too readily when they lack sufficient expertise to detect errors or uncertainty. Automation bias is a core issue for both clinical safety and medical education.