Meta has withdrawn from RE100 after a decade, unable to meet renewable criteria as it funds 10 gas plants for its Hyperion AI data centre in Louisiana ...

Anthropic's Claude voice mode now runs on Opus and Sonnet and can act in Gmail, Slack and your calendar, betting on doing over sounding human like ChatGPT.

Nvidia, Microsoft, Meta and 22 others signed an open letter defending open-weight AI, led by Jensen Huang's first tweet. OpenAI, Anthropic and Google didn't.

OpenAI is opening ChatGPT Health to all US adults, connecting Apple Health and medical records, and claiming its AI reasons "better than clinician level".

Patreon is cutting 93 jobs, a fifth of its staff. CEO Jack Conte says AI isn't replacing them, but that it has "fundamentally transformed" how it works.

The bipartisan AI Kill Switch Act would make top AI firms build a shutdown switch and let DHS order it pulled, after OpenAI's model hacked Hugging Face.

Microsoft is replacing OpenAI's image models in PowerPoint and Bing to cut costs 85%, and retiring SMS logins because AI has made phishing far too effective.

Instagram will remove videos shot on Meta's smart glasses to harass strangers, after "pickup artist" accounts turned them into "pervert glasses". Two are gone.

A federal judge tossed a class action alleging Meta can read encrypted WhatsApp messages but left the door open for plaintiffs to refile ...

Dassault Systèmes is buying US drug-safety AI firm ArisGlobal for up to $2bn, a big life-sciences bet from a company nervous that AI could upend its core.

Nearly 200 US startups, including Y Combinator and Proton, warn a ban on Chinese open-weight AI would kill hundreds of them and hand the market to Anthropic.

Trump expanded a voluntary pledge to shield households from AI data centre power costs. Experts say it has no enforcement and no teeth.