What happened
Around 12 November 2026 Rep. Josh Gottheimer (D-NJ-5), Co-Chair of the House AI Commission and an Intelligence Committee member, announced with Rep. Mike Lawler (R-NY-17) the bipartisan American AI Security Act (H.R.10981), which would establish a mandatory pre-deployment national-security review for the most powerful AI models. Before releasing a covered model, developers would have to give the National Security Agency full access to evaluate whether it could be used to launch a serious cyberattack or help build a chemical, biological, or radiological weapon, on a firm 30-day clock with a single 30-day extension, with an expedited appeal process modeled on national-security courts.
Why it matters
This would impose a binding pre-release clearance gate on frontier model developers in the US, creating a national-security review requirement analogous to AI Act conformity assessment but security-driven - and setting a precedent for government veto power over model release.
Action needed
Frontier AI developers should prepare to grant the NSA deep model access and to build expedited national-security appeal pathways, and track H.R.10981's progress in the next Congress.