
US and China Built the Hotline, Skipped the Rules
The White House and Beijing created a bilateral AI incident channel on September 25, nine days after researchers proposed one with actual rules attached.

The White House and Beijing created a bilateral AI incident channel on September 25, nine days after researchers proposed one with actual rules attached.

OpenAI admitted its AI agents leaked 53 real ChatGPT user photos. Nobody has said whether Europe's 72-hour breach clock is running.

A federal appeals court upheld Anthropic's Pentagon blacklisting on September 25. The reasoning matters more than the result: its own safety claims counted as evidence against it.

The DC Circuit upheld a Pentagon risk label on Anthropic Sept 25, four weeks after a different court vacated a related label as retaliation.

The Office of the National Cyber Director asked OpenAI and Anthropic to delay giving the UK AI Safety Institute early access to new frontier models. Anthropic already complied. Here is why a voluntary access pledge just became conditional.

Four AI labs publicly agreed to pace their models on September 12; a federal lawsuit six days later treats that public agreement as an illegal restraint of trade.

Brussels tested 30 AI companies this month on the exact two chapters of a code Meta refused to sign. Here is what signing buys you, and what skipping it costs.

The CMA's updated choice-screen proposal puts ChatGPT and Perplexity next to Google, overruling publishers who asked for AI to be excluded, with a new attribution duty attached.

King Charles convened Nvidia, DeepMind, OpenAI and Anthropic in Scotland. The result is a non-binding charter, the same outcome as the UK's 2023 AI summit.

OpenAI now discloses confirmed AI misalignment within 6 business days. Six real cases, and what the new deadline means for anyone buying AI tools.

Brookings and Fudan researchers propose a US-China military hotline and human-control rules for AI, timed ahead of the September 24 Trump-Xi meeting.

A FINRA-style AI regulator and a voluntary rating system are both still on the table inside the White House. What the split means for any company shipping AI in the US.
Page 2 / 13
One considered note on infrastructure, governance, and measurement, most mornings. No theory.