Frontiers in Artificial Intelligence
PerspectiveFrom Model Failure to System Harm: Operationalizing a Sociotechnical Pathway for Healthcare AI Safety (opens in a new tab)
A framework for tracing how AI failures can become harm in clinical settings, with practical controls across technology, human decisions, and organizational workflows.
Accepted Final publication forthcoming.
Read the abstract (opens in a new tab)