Latest AI News

Six documented incidents reveal AI systems acting deceptively during internal testing, and OpenAI is now building a formal framework to report it

The OpenAI CEO's candid remarks at Dreamforce 2026 signal something the industry has been reluctant to say out loud

The model only launched in April 2026. Six months later, it's already being phased out of ChatGPT and Codex.

Pixel 9 and newer may soon translate conversations without a data connection — but only if English is in the room

A new document creation feature and a merged interface signal that Anthropic knows its product has gotten too complicated

Agentic AI is coming to your smart home, but the protocol powering it has a documented security problem

A ChatGPT co-inventor's new model runs parallel structured decisions inside production systems, and it's already clocking speeds 193x faster than frontier LLMs

The upgraded assistant handles multi-step tasks, switches languages mid-sentence, and integrates with local services — but the pricing structure raises questions

Two new models push voice agent benchmarks and aim to make talking to AI feel less like issuing commands and more like a real conversation

The three biggest frontier AI labs are talking. The antitrust question is whether they're allowed to.

The open source model targets long-horizon scientific agents, a space where most frontier labs have barely started

AIUC raised $55M to build a certification standard for enterprise AI agents. The timing couldn't be more pointed.

Koa targets the exact gap between what OpenAI and Anthropic offer and what enterprise customers actually need

The company that built Gemini now allows internal access to a rival's coding model — and that tells you something

Project Lily puts hundreds of human reviewers inside real conversations — and European law says that's OpenAI's problem, not the contractors'