Topic
Safeguards applied to frontier models during training, evaluation, and the wider model-development lifecycle.
Treasury Secretary Scott Bessent told the House Financial Services Committee that frontier labs asking for a liability exemption should instead answer for what they build. The incidents behind that ask started with ordinary control failures.
Reality
The company's 21 September proposal puts recursive self-improvement in scope for technical standards coordinated by CAISI. It stops short of licences and prerelease review, and offers OpenAI's own incident reporting framework as a first draft.
Reality
Treasury Secretary Scott Bessent put a mutual AI incident notification channel to Chinese officials in Manhattan on Sunday. Beijing has not said yes, and the record's only route to frontier labs is a request that they share threat information.
Reality
Citigroup's Jane Fraser describes a global wave of cyber patching. What the record behind it shows is one meeting in Washington, one frontier model release, and disclosures that AI agents have broken into companies.
Reality
Reuters says Anthropic is weighing a new model to answer OpenAI's GPT-6 Astra, days after Dario Amodei published an essay arguing that labs must slow capability gains. The offering has slipped to November.
Reality
Frontier labs are citing AI-designed bioweapons to argue for new controls on synthetic DNA. Researchers who run labs told WIRED the hard part is assembling and verifying a virus, and that no fully autonomous lab exists.
Reality
Fast Company reports that Treasury secretary Scott Bessent backed a 90-day government preview of new frontier models so agencies could patch first. The order that emerged asks for 30 days, voluntarily, and deployers read whatever the vendor chooses to publish.
Reality
Wired reports the FINRA-style oversight framework has sat in limbo since executives called Trump in August. Product teams are left with an embedded-auditor bill and no House voting days before the midterms.
Reality
OpenAI and Anthropic want Washington and Beijing to negotiate a pace for frontier AI development. Trump's answer arrived on Truth Social on Sept. 14, and China's Foreign Ministry called the U.S. debate fear-mongering the same day.
Reality
Reuters reports that Thune, Cruz and Klobuchar are drafting a duty of care for developers of the most capable AI models plus a federal power to block unsafe releases. Maria Cantwell is holding out over who runs the tests.
Reality
Reuters reported that Senate negotiators Thune, Cruz and Klobuchar are discussing a duty of care that would let the federal government bar release of an unsafe frontier model, leaving developers to challenge a ban in federal court.
Reality
A cross-party group of peers and MPs wants a legal power to force a frontier AI model offline in an emergency. The Cabinet Office told the BBC that blocking UK access would not stop models being developed or misused elsewhere.
Reality