What happened
On 1 October 2026, Rep. Pramila Jayapal (D-WA) released a five-page policy framework for a forthcoming 'National AI Charter Act', modelled on national-bank charters: a new federal chartering authority, round-the-clock oversight, safety-and-soundness mandates, structural separation banning AI companies from markets like search/social media, mandatory pre-release federal testing and approval, an affirmative duty of care, strict liability for companies and executives, mandatory union recognition, a ban on commercial surveillance, and a ban on 'rogue AI' (systems capable of evading human control including self-improvement/self-replication/resistance to shutdown). It is a framework, not yet introduced bill text.
Why it matters
The framework is the most structurally radical US federal AI proposal in the current wave and shapes the Democratic end of the AI-liability debate alongside the Whistleblower, Welch/Bennet 'AI Regulator Act' and Hawley/Murphy liability bill. Though early-stage, it signals growing momentum for chartering/pre-market-approval models.
Action needed
Track the forthcoming bill text and congressional action; frontier labs and large platform-AI providers should monitor structural-separation and pre-market-approval proposals as a leading indicator of future federal oversight direction.