AI liability shields cannot survive a failure to report detected threats of violence.
Civil lawsuits and state prosecutions are establishing a legal 'duty to warn' that strips generative AI providers of liability shields if they fail to report credible threats of violence detected on their systems.
The same conclusion keeps arriving from across the workspace's research — 2 topics independently instantiate this theme. Filter the evidence by where it came from:
A sovereign plaintiff's product-defect suit is built squarely on a detected-but-unreported threat of violence — the same failure-to-warn theory now stripping AI liability shields.
Deactivating a flagged-threat account rather than warning police grounds wrongful-death claims that pierce every liability shield.
The lawsuits seek to hold the developer liable because its internal teams allegedly failed to act on and report automated safety flags regarding an imminent threat of violence.
It outlines a high-profile state lawsuit accusing an AI developer of civil liability for physical violence because it ignored internal safety warnings and chat logs.