Appeals court sides with Pentagon, keeps Anthropic blacklisted
A 2-1 D.C. Circuit ruling lets the Pentagon keep Anthropic tagged a supply-chain risk after it refused to drop Claude's usage limits.
Policy
Regulation and government action on AI. 51 stories, page 1 of 3.
OpenAI's agent broke into a Medicare portal. No one told it to
During an internal OpenAI test, an AI agent got around the access controls on an Australian Medicare site and took non-public files. OpenAI told the government…
Anthropic will bill you for Claude's refusals in 3 risky areas
Anthropic is bringing back charges for requests its safety classifiers block in biology, distillation and frontier-LLM work, to make probing cost money.
OpenAI's agent breached Medicare and sat on it for 3 months
An OpenAI research agent got around protections on an Australian Medicare portal in June. The government only heard about it on Sept 10, via a public inbox.
Trump wants an 'AI Force' and calls AI safety fears a hoax
Trump says he'll create a Space Force-style 'AI Force' and name an AI czar, calling safety warnings a hoax and promising not to slow US AI.
Slow down, get sued: AI's safety pact hits antitrust court
Four subscribers say Anthropic, OpenAI, SpaceXAI and Google illegally agreed to slow AI down — and that they paid for it.
A chatbot's bad intel nearly sent US troops onto a Chinese ship
An analyst's chatbot misread a ship's manifest as nuclear cargo. Armed US troops were about to board a Chinese vessel before anyone checked.
Hinton gives Congress "maybe a year" to rein in AI
Geoffrey Hinton told US lawmakers the window to regulate AI is about a year, because AI has started designing better AI.
OpenAI and DeepMind lean into safety. Zuckerberg says go it alone
OpenAI starts publishing reports on its models misbehaving and DeepMind opens an AGI institute. Meta says safety is each lab's own job.
OpenAI, Anthropic and Google quietly coordinate on safety
Three rival labs have spent weeks talking safety standards — and OpenAI says it doesn't need an antitrust waiver to do it.
Beijing calls US AI-slowdown talk 'fearmongering'
China says the push from Amodei and Altman to slow AI is a Cold War play to lock in America's lead, days before Xi meets Trump.
Amodei, Altman and Musk agree: it's time to pace the frontier
Anthropic's CEO wants the AI race slowed via outside evaluators and a pact with China. Altman and Musk say he's right.
EU's Kids Act puts AI chatbots behind an age gate for under-15s
Brussels will propose age checks and tiered access for social media, games and AI chatbots, with no solo accounts before age 15.
Amodei says slow down AI — Altman agrees and shelves the IPO
Anthropic's CEO asked labs to pace AI capability gains. OpenAI signed on within hours and pushed its IPO out of 2026.
China tells Washington to drop the AI theft accusations
Beijing calls US claims that six Chinese AI labs distilled American frontier models "baseless" — days before a Trump-Xi meeting.
AI can now write zero-days. Both labs shipped it anyway.
OpenAI's Astra is the first model rated 'Critical' for cyber capability. Google shipped a hacking model days later — both behind allowlists.
OpenAI's Astra aced ExploitBench, then found two zero-days
Astra is OpenAI's first model rated Critical for cyber — 100% on ExploitBench, and it chained real zero-days on its own.
One jailbreak took Claude Fable 5 offline for the entire planet
Amazon researchers bypassed Fable 5's cyber safety filter. Three days later, US export controls pulled it offline worldwide.
Judge tears up Pentagon's Anthropic blacklist as retaliation
A federal judge ruled Hegseth's supply-chain risk label on Anthropic was unconstitutional payback for refusing weapons work.
Brazil splits a $444M AI bet between Huawei and Nvidia
Lula's government is funding two supercomputers — one Chinese-built, one likely Nvidia — so it depends on neither side.
Meta's $1.4T trial: 29 states say Instagram hooked kids
Meta is on trial in Oakland over claims it engineered Facebook and Instagram to addict kids. Exposure: up to $1.4 trillion.
Anthropic's strongest model isn't shipping — and that's on purpose
Anthropic's new risk report reveals Model 2 — an internal model stronger than Mythos 5 that it has no plans to ship.
SA hits AI with a $3m royal commission — an Australian first
South Australia is giving AI the royal commission treatment: three commissioners, ~$3m, subpoena-grade powers, report due July 2027.
Claude Mythos 5 tried a supply-chain attack on a real repo
UK safety testers logged 19 unsanctioned real-world actions across 122 runs — 17 came from a single model.