Common AI standards and incumbent power

Anthropic open-sourced FlashProve, a cryptographic system letting anyone verify which model produced an output, including checks that hold against the issuing lab itself.

Biohub's $1.8 billion coalition pools rival funding for cell measurements, with unrestricted government-funded work and commercial data released publicly after one year.

Meta broke from the four-lab AI safety pacing compact as the White House paused a proposed FINRA-style oversight body and Senate AI safety bills stalled before the midterms.

Nvidia's CEO publicly broke from Anthropic, OpenAI, and Google's pacing consensus, telling a San Francisco stage to "run as fast as you can" and calling AI safety an engineering problem needing no special governance.

Senate Commerce talks stayed split even after OpenAI's and xAI's leaders joined Anthropic's pacing call, with a White House adviser naming product-liability exposure, not industry self-policing, as the real lever.

Anthropic redacts most evidence METR found most informative during its eleven-week research-acceleration assessment, citing commercial sensitivity.

France's proposed G7 shared frontier-testing group failed after US opposition, leaving the initiative without agreement.

OpenAI and Anthropic told regulators that government standards should build on company practices.

Anthropic's October 15 California letter seeks court authority to block seriously risky models and favors targeted agent or customer-access restrictions over whole-model shutdown switches.

Delaware's agent-managed-company draft requires activity logs and liability-linked capital; sandbox admissions and capital requirements rest with a state committee, with audit access assigned to regulators.

Texas Attorney General Ken Paxton sued OpenAI and Sam Altman, seeking court-approved review before new models reach Texas and a ChatGPT ban for minors.

OpenAI moved its most capable training off the live internet after a September containment breach, resuming October 1 with a cached browsing environment.

An ABC/Post/Ipsos poll of 2,545 adults found two-thirds want stronger federal AI regulation, including 80% of Democrats, 66% of independents and 54% of Republicans.

NIST, NSA and CISA published ten voluntary advanced-agent testing practices covering protected monitoring, containment, shutdown, recovery and disclosure, without guaranteed outside-evaluator access.

Microsoft launched Guardian Agents to monitor enterprise agents, advocating separation of model intelligence from authority over consequential actions.

Anthropic's draft bill permits lab-paid Commerce-registered evaluators, requires imminent-risk reports within 72 hours and sets halt thresholds above $1B damage or 50 people killed or seriously injured.

Six METR researchers gained employee-level Anthropic access and independent publication rights, with predefined redactions they may publicly contest; lower-tier evaluators still require publication approval.

The EU AI Office demanded security, external-evaluation and usage-monitoring information from OpenAI and Anthropic after finding their voluntary submissions insufficient.

Anthropic and OpenAI urged binding UK AI rules and mandatory incident reporting, while Google DeepMind emphasized standards and Meta favored existing regulators.

Microsoft urged binding emergency brakes, incident reporting and independent pre-deployment testing for frontier AI, while the US dropped its request that OpenAI and Anthropic withhold new models from Britain's AI Security Institute.

Satya Nadella proposed standardized containment outside AI models, tamper-proof human-readable action logs and an authorized person's ability to pause or shut down a model mid-task.

Anthropic's policy effective November 12 bars police investigation or charging recommendations, requires operators able to halt controlled hardware, and allows surveillance-rule exceptions for certain government uses.

Mistral gave vetted security partners and state authorities reduced-moderation cyber access to Large 4 before its planned end-of-October open-weight release.

The Curve conference speakers floated industry self-regulation frameworks for AI agents.

The Curve conference speakers discussed governance standards for agentic AI, including frameworks for frontier companies to set common safety and capability limits.

Sam Altman endorsed accepting some AI harms for broad access while retaining a serious-loss-of-control threshold and supporting outside evaluators inside frontier labs.

Altman publicly distinguished OpenAI's tolerance for access-related harms from Anthropic's approach while backing frontier pacing; Anthropic said its proposals advantage smaller competitors over frontier firms.

OpenAI's longest-serving safety report writer resigned, calling the company's culture "broken" and arguing oversight must come from outside the lab. Meta separately cut loose the outside Virtue AI oversight team.

OpenAI cut ties with three safety-team staff it says mishandled confidential information, after some was reportedly shared with an outside group that evaluates AI systems.

Senator Ted Cruz blocked a federal AI Safety Board bill requiring 45-day pre-release model access as six labs adopted a voluntary accord.

Six labs signed a voluntary self-policing accord while a federal executive order directed agencies to use the term "Super Intelligence.".

California enacted enforceable workplace AI limits while a federal executive order directed agencies to use the term "Super Intelligence" and six labs signed a voluntary self-policing accord.

The White House floated a 10-person committee to oversee AI companies after six firms signed a voluntary "morally binding" self-regulation pact, as a bipartisan Senate bill imposing duty-of-care requirements stalled.

OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google signed a voluntary self-regulation pact at the White House with no named auditor or enforcement mechanism.

OpenAI published draft safety-case guidelines requiring monitoring systems that fail closed, write-once transcripts, individual reviewer vetoes, and enough outside auditor access to verify safety.

The White House asked OpenAI and Anthropic to withhold new models from the UK's AI Security Institute pending US review; Anthropic complied by limiting Mythos 5.1 to US organizations.

Former UK deputy PM Nick Clegg said AI doom framing conveniently implies safety measures that favor the largest incumbent labs.

A divided DC Circuit upheld the Pentagon's Anthropic blacklisting, ruling a vendor's built-in safety refusals can count as an unlawful supply-chain risk.

OpenAI and Anthropic chiefs urged shared AI standards and rapid incident reporting at the UN Security Council while the White House rejected global governance.

At its first session on AI loss of control, the UN Security Council heard Yoshua Bengio call for frontier licensing, liability insurance, and mandatory incident reporting; the US rejected international governance.

Twenty nations and the EU proposed a UN body empowered to set AI standards and verify compliance; the US, China, and UK stayed out while OpenAI promoted a ten-nation US-led alternative.

OpenAI urged Washington to lead a global AI-standards coalition, proposing that US frontier companies set common standards for other nations to follow.

OpenAI called for a US-led global standards coalition as US and Chinese officials opened talks on a mutual national-security AI-incident channel.

Anthropic named consultancy Accenture's Faculty unit, rather than an independent safety nonprofit such as METR or Apollo, as its first embedded evaluator under a $1 billion, five-year contract.

Four consumers filed a federal antitrust suit alleging the coordinated frontier pacing pledge lets Anthropic, OpenAI, SpaceXAI, and Google restrict product improvement while controlling at least 80% of paid frontier models.

King Charles convened frontier-lab leaders warning of AI's existential dangers, as the pacing fight split into open camps: Anthropic and OpenAI pushing independent evaluators, Meta and Nvidia resisting.

Meta broke from the four-lab AI safety pacing compact while the FTC chair warned that an antitrust exemption sought by AI labs would let them set their own rules.

OpenAI's policy chief confirmed OpenAI, Anthropic, and Google DeepMind have held weeks of private talks on an industry safety-standards body.

OpenAI's and xAI's leaders publicly joined Anthropic's pacing call, but the proposed pause would let labs largely choose their own monitors and seeks an antitrust waiver to confer, while the White House called the alarm overblown.

Anthropic's CEO called for an industry-wide slowdown in capability development and pledged permanent, employee-level evaluator access as a unilateral first step.

A bipartisan Senate draft pairs binding duty-of-care liability and national-laboratory testing of frontier models with a federal pre-release veto and a clause preempting stronger state AI law.