ChatGPT crosses the EU's 45 million-user threshold

The European Commission designated ChatGPT as a Very Large Online Search Engine after it reported at least 45 million average monthly users in the EU. ChatGPT, Reddit, and Roblox have until January 2027 to meet additional Digital Services Act obligations for assessing and mitigating systemic risks.

Frontier-model cyber risk reaches the G20 agenda

The Guardian

Bank of England governor Andrew Bailey, writing as chair of the Financial Stability Board, warned finance ministers and central bankers that advanced models could increase cyber risks to the global financial system. The intervention treats model autonomy as a financial-stability concern rather than only an operational security problem.

Texas freezes state funding for more Flock cameras

The Verge

Governor Greg Abbott froze additional state spending on Flock's AI surveillance cameras after Texas had already spent more than $30 million, largely through a fee added to insurance policies. Existing cameras and local deployments remain in place unless officials take further action.

OpenAI reportedly tests payment on completed tasks

The Information via Techmeme

OpenAI has reportedly offered some major customers the option to pay when its AI completes tasks, while Salesforce and other providers test similar pricing. Such contracts need a definition of completion and acceptable quality, along with rules for retries, reversals, and human review.

Music publishers sue Anthropic over training works

The Guardian

Sony Music Publishing and Warner Chappell are seeking damages over Anthropic's alleged misuse of tens of thousands of copyrighted songs to train Claude models. The lawsuit puts the disputed acquisition of training rights into a potentially multibillion-dollar case.

Google's studio licensing talks reportedly stall

Los Angeles Times via Techmeme

Google reportedly approached Disney, Universal, Warner Bros. Discovery, and other studios about licensing characters and film libraries for AI tools. Legal and union concerns stalled the talks, and the report doesn't describe completed deals.

NHS watchdog finds errors in AI-generated consultation records

The Guardian

Healthwatch England found that AI scribes can record the wrong drugs and diagnoses, including errors that patients caught after clinicians missed them. The cases put verification at the center of a tool used to create consequential medical records.

AI labs put more capital and attention into medicine

Axios

Anthropic, Nvidia, Eli Lilly, and Isomorphic Labs are among the companies investing in biology, drug discovery, and clinical applications. The near-term record includes useful clinical work as well as higher costs, security concerns, and benefits that may take years to establish.