Ex-OpenAI Policy Lead Reverses Stance: No Liability Shield for AI Labs, Only Rebuttable Presumption

deanwball · x · 2026-08-28

Dean W. Ball publicly details his shift on tort liability for AI developers: he stopped supporting the 'liability shield in exchange for rigorous independent technical assessment' trade over a year ago. He now supports at most a rebuttable presumption of due care if a lab has undergone independent verification by a government-accredited body for the risks relevant to the alleged harm.

He notes independent technical assessment has moved from fringe to mainstream, with Tyler Cowen endorsing a version of it, and stresses these are his personal opinions, not OpenAI's. Philip Wallach had questioned whether such proposals amount to special pleading for AI companies.

Original post →

More from Safety

Safety channel →