The Free Will Show Episode 114: Responsibility for Consequences and the AI Responsibility Gap with Huzeyfe Demirtas
Mar 2, 2026
Huzeyfe Demirtas, a Harvard postdoc studying moral responsibility and AI ethics. He discusses responsibility for consequences using vivid examples. He examines resultant moral luck and critiques the degree–scope reply. He explores the so-called AI responsibility gap and argues it may be inevitable yet manageable, with practical legal and regulatory implications.
AI Snips
Chapters
Transcript
Episode notes
Rock Throwing Demonstrates Consequences Beyond Control
- Throwing a rock example: agent wills to throw, releases rock, then consequences follow beyond agent control.
- Demirtas uses this to define 'consequences' as whatever follows the mental act of willing or attempting.
Why Resultant Moral Luck Is Problematic
- Resultant moral luck is when outcomes beyond control affect blameworthiness, e.g., a gust of wind changing whether the rock harms someone.
- Demirtas rejects it because consequences after the act don't affect control or the quality of will that grounds blame.
Problems With The Degree Scope Response
- The degree-scope response says degree of blame stays same while scope (what you're responsible for) varies with outcomes.
- Demirtas argues this view is unclear: 'responsible for' either means causal, duty, or part of the basis of blame, and each interpretation undermines the view.