California lawmakers moved a revised AI transparency package toward a floor vote, requiring developers above a compute threshold to publish safety summaries and incident reporting channels for critical-use deployments. Trade groups in San Francisco and San Jose said the amendments narrow scope compared with earlier drafts but still risk patchwork rules if other states copy the template.
The White House Office of Science and Technology Policy told reporters that federal frameworks remain the preferred path, though progress in a divided Congress has slowed ahead of midterm campaigning. Consumer advocates argued state action is necessary after high-profile failures in automated hiring and healthcare chat interfaces.
Open-source coalitions asked for carve-outs for academic and nonprofit releases, while enterprise vendors sought harmonized audit formats to avoid 50-state checklists. Legal scholars at Stanford and Berkeley scheduled a joint briefing on constitutional preemption questions.
European diplomats in Washington watched closely — Brussels officials said US state experiments could align with EU AI Act documentation norms or complicate transatlantic product launches. Track US politics and tech on wop360.com/united-states.
