The immediate ban covers the US, Japan, and five other markets, and also blocks third-party ads that route users to ByteDance apps.

A voluntary AI safety pact leaves companies to decide when risk is real; the ex-official who faced H5N1 says predefined triggers are the missing plan.

Apple's Mac desktops are the most sought-after computers of the year as agentic AI takes off. WIRED's guide breaks down the Mac Mini, Mac Studio, and iMac by use case.

Researchers found leading AI models can produce realistic election-fraud images and documents, exposing a widening gap between AI capabilities and election safeguards.

Anthropic's policy overhaul bans sustained cruelty to Claude, but the consequential changes target election interference, weapons development, and surveillance.

Times Square line shows Insta360 winning creators as DJI faces U.S. blocks.

A he-said-she-said over AI safety: the fired researchers claim they were punished for putting safety first, while OpenAI says sensitive information was mishandled. The outcome could shape how AI labs treat internal dissent.

Meta's safety-first pause on its A.I. agent app collapses under competitive pressure, signaling a speed-over-caution shift in the A.I. race.

South Korea's RAIBO2 quadruped finished 26.2 hilly miles without recharging, and its energy cost now beats the human benchmark.

The Australian AI data center operator shelved a historic listing, signaling how tough the window is for capital-hungry AI infrastructure plays.

Samuel King's generative AI produced genetic blueprints for microscopic viruses in 2025, and the MIT Technology Review Innovators Under 35 honoree says AI-generated life may be the next step.

The AI lab is arming defenders with its strongest models and human engineers, betting AI will flip the security advantage in two years.