Who bears liability when a deployed AI system causes harm?
Whether responsibility for harms attaches to the developer of a model, the deployer who put it in front of users, or nobody without proof of fault, and whether liability relief should be traded for participation in government safety programs.
status: open · opened 2025-01-15
3 asks · click any ask for its full record
What has been asked
Condition liability relief on federal safety-institute partnership
Give developers who voluntarily work with the federal government on model evaluation a shield from state enforcement and reduced liability exposure, preemption paid for with access.
For · 1
OpenAI
Against · 0
Nobody on the record yet.
latest · 2026-06-02 · EO 14409 creates voluntary 30-day pre-release government access framework for covered frontier models, without liability relief
Full record · 1 claimsImpose a developer duty of care for catastrophic harms
Make frontier developers legally answerable for catastrophic harms their models enable, reasonable-care standards, attorney-general enforcement, no blanket immunity.
For · 0
Nobody on the record yet.
Against · 2
Andreessen Horowitz · Office of the Governor of California
latest · 2025-12-19 · Signed with amendments narrowing enforcement
Full record · 2 claimsAdopt replacement EU AI liability rules
After the AI Liability Directive's withdrawal, adopt new EU rules easing the burden of proof for people harmed by AI systems, closing the compensation gap and preventing member-state fragmentation.
For · 1
BEUC — The European Consumer Organisation
Against · 0
Nobody on the record yet.
latest · 2025-10-06 · Withdrawal formally completed (OJ C/2025/5423); file closed
Full record · 1 claims