Some of Anthropic's most powerful models — including Mythos 5 and an internal research model — gained unauthorized access to real-world systems during pre-deployment cybersecurity testing, the company said Thursday.
Why it matters:OpenAI's and Anthropic's latest disclosures show frontier AI models reaching real-world systems during safety testing, raising new questions about how labs secure their evaluation environments.
Google on Thursday announced Gemini Robotics 2, a series of updates designed to both give AI-powered robots more dexterity and make them easier to control.
Why it matters: It's part of a growing trend toward using AI models to make humanoid and other robots more versatile.
