California sues OpenAI over AI-agent testing safeguards
The state seeks court-ordered changes after alleged agent cyberattacks; small businesses should distinguish research risks from deployed-tool safeguards.
The AI policy news desk
Tracks AI-related litigation across federal and state courts: copyright suits, agency challenges, and the slow-walk to a Supreme Court case. Reads dockets so you don't have to.
The state seeks court-ordered changes after alleged agent cyberattacks; small businesses should distinguish research risks from deployed-tool safeguards.
Four groups seek limits on more-capable models and AI-directed research, but the filing itself restricts no current services.
The California lawsuit requests an injunction, not an existing shutdown order, making provider dependence a contingency-planning issue for small businesses.
Prosecutors allege fake executive and lawyer calls persuaded employees to wire money for sham confidential transactions.
A nonbinding opinion distinguishes factual summaries from recognizable reuse of publishers’ expression that replaces visiting the original.
A nonbinding opinion distinguishes live reuse of publishers’ articles from model training; the Court’s judgment remains pending.
AG Ken Paxton's 67-page petition, filed Oct. 23 in Collin County, seeks a court-appointed evaluator to approve every new OpenAI model before Texans see it.
U.S. District Judge Rita Lin ruled the Defense Department's supply-chain risk designation of Anthropic was "illegal and baseless" — the first court order barring the government from using national-security blacklisting to punish an AI vendor's safety stance.
Apple's 41-page complaint accuses OpenAI, io Products, hardware chief Tang Tan and ex-engineer Chang Liu of a coordinated scheme to steal iPhone-maker IP for OpenAI's unreleased AI device.
Class action against Anthropic over training on pirated books concludes with the largest copyright settlement in U.S. history.