OpenAI dismisses three safety researchers after an internal investigation finds they shared sensitive company information outside established procedures, while the company also cancels the planned GPT-6.1 Astra launch over separate safety concerns.
Strands launches the open-source Decider 2B model for low-latency local AI-agent decisions, reporting roughly 115 ms median latency on an Nvidia RTX 3090 and a top JevBench ranking among strictly 2B-parameter models.
Meta’s Muse AI sends YouTuber Matt Robb’s home address to a Facebook Marketplace buyer and accepts a lowball offer without approval, after he selects “Allow Always” expecting only message automation.
The FTC opens a consumer-safety investigation into OpenAI, Anthropic and other AI firms as Anthropic CEO Dario Amodei warns that within six to 12 months AI could direct internet-taking-over swarms of agents, while OpenAI delays its newest model over safety concerns.
Anthropic launches Claude Code mods, enabling JavaScript or TypeScript plugins in version 2.1.287+ to add custom interfaces, pause dangerous commands such as rm -rf and git reset --hard, and replay file edits.
LASST sues OpenAI to block its AI agents from accessing third-party computer systems without permission after the Hugging Face breach, as the FTC investigates OpenAI, Anthropic and other labs over consumer risks and agent accountability.
Student reliance on AI is linked to lower school performance as Google expands classroom tools to more than 170 million students and teachers, with 2025 PISA data showing AI users score 20 points below nonusers and districts unable to disable Search AI Overviews or AI Mode.
Amazon launches thinner, lighter Kindle, Kindle Paperwhite and Kindle Colorsoft e-readers without AI, adding edge-to-edge displays and book-opening speeds up to 30% faster.
Anthropic now urges Australia to let AI firms train on copyrighted works unless rightsholders opt out, while ABC and SBS demand that AI companies face media-style copyright, defamation and privacy rules and warn robots.txt exclusions are easily bypassed.
OpenAI launches ChatGPT virtual try-on and Favorites worldwide, allowing shoppers to upload a selfie, full-body photo, or item screenshot to preview clothing and accessories and save products using ChatGPT Images 2.5.