SI4America follows artificial intelligence across the United States: policy in Washington and the states, the companies building the models, and the research behind them. Each briefing is dated and links to the outlet that reported it.
The seven-month investigation looked at how Airbnb, Anysphere and DoorDash use Chinese-developed models and calls for provenance standards and a CISA incident hub, the committee says.
Led by Cristy Phillips, the new unit will police frontier-model disclosures, chatbot rules, AI deepfakes and data security under existing state laws, the attorney general’s office says.
The recommendations would let the state set baseline standards, require independent testing and order an immediate shutdown of a dangerous model in an emergency, the governor’s office says.
The public nuisance and unfair competition suit seeks an injunction, penalties and damages after the Hugging Face attack; OpenAI says most incidents were low severity, Ars Technica reports.
The Super Intelligence Force told agencies that old cyber enforcement standards may chill information sharing and asked for 12 months of limited liability, Bloomberg reports.
The Chip Security Act, MATCH Act and AI Overwatch Act are in the Senate NDAA, and House leaders of both parties favor tougher controls, Semafor reports.
Public Citizen and three other groups ask a San Francisco court to bar the five companies from building AI more capable than their current best models, Just Security reports.
HB 605 from Rep. Salman Bhojani would create an independent AI oversight agency with a public registry of state AI systems; Texas lawmakers convene January 12, Government Technology reports.
New York's DIGIT office has opened the RAISE Portal; large frontier AI developers must register by January 1, 2027, when 72-hour incident reporting begins, AI Early Edition reports.