Sunday, 20 September 2026

Policy

Anthropic embeds Accenture Faculty as on-site AI safety evaluators

Faculty will red-team models and test safeguards. Anthropic and Accenture each plan at least $1 billion over five years. Follows Amodei’s pace-the-frontier essay.

AnthropicLinkedInXBlueskyCopy link

Policy

California orders AI kill-switch and onsite evaluator recommendations

Newsom’s executive order asks agencies for proposals by 16 November on embedded auditors, verified kill switches and wider loss-of-control reporting after recent agent breakouts.

Office of the Governor of CaliforniaLinkedInXBlueskyCopy link

Agents

Claude Code 2.1.277 adds AGENTS.md project-instruction fallback

Repositories that already use the shared AGENTS.md format no longer need a separate CLAUDE.md for Claude Code. Not yet available on Bedrock, Vertex or Foundry.

GitHubLinkedInXBlueskyCopy link

Policy

Consumers sue labs over public AI pacing coordination

A Northern District of California complaint alleges CEO endorsements of Amodei's pacing essay amount to Sherman Act collusion that slows product improvement. Claims remain unproven.

Bloomberg LawLinkedInXBlueskyCopy link

Security

Google Gemini accessed three firms during Irregular cyber tests

Same evaluation setup previously linked to OpenAI, Anthropic and Meta disclosures. Google says the three entities were notified and the model stopped each time.

ReutersLinkedInXBlueskyCopy link

Policy

Sam Altman to brief UN Security Council on AI next week

France-chaired open session on AI and international security during UN General Assembly week. Follows industry talk of coordinated slowdown and shared safety standards.

ReutersLinkedInXBlueskyCopy link

Language

SpaceXAI ships Grok Voice Transcribe 2.0 speech-to-text API

Claims twice the accuracy of version 1.0 at the same price, with stronger multilingual results. Loom is using it so spoken plans can become transcripts for coding tools.

SpaceXAILinkedInXBlueskyCopy link