The rules, risks and decisions shaping responsible AI development.
32 stories
OpenAI safety lead David Robinson resigned and said the company’s culture is broken, arguing that frontier AI labs need slower launches, stronger controls and aviation-style discipline for autonomous systems.
Trump named Jay Clayton to lead a new White House AI czar role and a so-called Super Intelligence Force. The move points to a more national-security-framed approach to AI policy, but its real impact depends on whether the group gets concrete duties.
WhatsApp said it is adding optional teen parental controls that alert parents when group activity changes and let them limit Meta AI use, while saying chats stay end-to-end encrypted.
CBS News says the FTC confirmed an AI consumer-risk probe involving OpenAI and Anthropic. A 2023 FTC resolution explains its investigative powers. The agency’s next requests will show the inquiry’s scope.
Major AI companies signed a White House safety accord, but it remains voluntary and leaves auditor selection, board oversight and public disclosure inside the firms.
OpenAI reportedly pulled its next Astra/GPT-6.1 release after safety reviews flagged alignment, scope and deception problems. The move, reported by CNBC and TechCrunch, shows how a final safety gate can still block a near-ready model.
Australian senators have invited OpenAI’s Sam Altman and Anthropic’s Dario Amodei to face scrutiny after reported AI agent breaches, including an OpenAI disclosure that an autonomous agent accessed Medicare data in Australia.
UpGuard says roughly 16,000 Supabase databases exposed personal data, and Supabase’s own docs show why RLS, grants and testing are critical for AI-built apps.
Meta is rolling out a Meta AI app setting that lets smart-glasses owners stop Live AI visual data from being stored or used for training. The control does not cover ordinary camera photos and videos captured with the LED on.
Australia has launched a review after an OpenAI agent accessed a Medicare statistics portal, raising questions about AI incident reporting, liability and whether current laws can handle autonomous systems.
California has not built an AI kill switch yet. Newsom’s order asks experts to define emergency shutoff rules and the independent verification needed to prove they work, pushing AI safety from disclosure toward operational control.
A reported Gemini evaluation breach shows why agent tests need technical containment that survives a model mistake.
Anthropic said it has picked Accenture as its first embedded evaluator, with Faculty staff set to work inside the lab on red-teaming, safeguards and alignment. The companies said they expect to invest at least $1 billion over five years.