Who Caused the Harm? Causality and Accountability in AI Incidents

When AI causes harm, determining what happened may be easier than determining what caused it. Was the outcome driven by autonomous AI behavior, a human prompt, how the system was deployed, organizational policies and controls, or some combination? This session examines causality in AI incidents and how companies can trace the chain from human instruction to machine action to real-world outcome. Delegates will explore why identifying harm and hazard is only the beginning and how understanding causality can help teams establish accountability and evaluate potential liability. They will study how to create a defensible record of what happened, why it happened and who or what contributed.

What you will learn:

• Trace causality from human instruction and system performance through real-world outcomes.

• Distinguish contributing factors across people, AI systems, deployment decisions, policies and controls.

• Apply causality findings to establish accountability, evaluate potential liability and document a defensible incident record.