The Maivia Gazette

Verified AI news, every morning

Business & Policy

House commerce chair signals no vote on the FRONTIER Act this year as evaluators say Amodei's plan needs legal backing

Brett Guthrie refuses to commit to a timeline for the bipartisan AI safety bill, while third-party evaluators and security veterans question how independent embedded auditors can be.

A long empty marble corridor with closed doors leads toward a wall calendar in slanted afternoon light.
AI-generated illustration, not event photography. The motion is AI-generated from the still.

House Energy and Commerce Chairman Brett Guthrie said Wednesday that he will not commit to a timeline for a committee vote on the FRONTIER Act and hinted there likely will not be one this year. 'I'm not going to say that the bill is going to move,' the Kentucky Republican said at a Politico event in Arlington, Virginia. 'It's really complicated, and I wouldn't want to do something in a lame duck session to do it quickly and not get it right.' The bipartisan bill, co-sponsored by Reps. Jay Obernolte and Lori Trahan, is widely regarded as Congress's best near-term chance to set AI safety guardrails and has backing from OpenAI, Anthropic and policy groups. Obernolte has pushed for a committee vote in November. The delay matters because the labs' own plan leans on outside enforcement. Anthropic CEO Dario Amodei proposed embedding third-party evaluators inside frontier companies with power to report incidents and assess alignment, committing to give groups such as METR and Redwood Research unprecedented access, and Sam Altman said OpenAI would follow. Evaluators who spoke to TechCrunch welcomed the idea but said details must be settled, ideally in legislation, before they can know whether they will be independent watchdogs or vendors on the companies' terms. Deeper access matters as models get better at recognizing when they are being evaluated. Security veterans offered a blunter critique. Katie Moussouris of Luta Security told TechCrunch the proposal 'seems like they're outsourcing,' arguing labs should first apply basics such as logs and permissions as rigorously to AI systems as they do to human users.

Sources

  1. The RecordKey lawmaker suggests action on AI safety legislation will wait until 2027Published · fetched
  2. TechCrunchAnthropic and OpenAI want to embed safety evaluators. Will they really be independent? | TechCrunchPublished · fetched
  3. TechCrunchAI labs want in-house auditors — but maybe they should shut the front door first | TechCrunchPublished · fetched

Also in this edition