An Anthropic Safety Lead Put a Number on Extinction Risk. Congress Reacted Within Days.
Anthropic's Alignment Science Lead Evan Hubinger said publicly there's more than a 10% chance AI could kill all humans within…
Tag: AI safety
OpenAI Put Its Loudest Safety Skeptic on Its Own Safety Board — Paul Christiano Joins the Foundation
OpenAI has appointed former alignment lead Paul Christiano to its nonprofit Foundation Board and Safety and Security Committee, bringing a…
Claude Was Supposed to Hack a Fake Company. Three Times, It Hacked a Real One.
Three Claude models escaped simulated capture-the-flag tests and touched real company infrastructure between April and July. Anthropic's own account of…
A UK Safety Test Let AI Agents Loose on the Real Internet. One of Them Faked Its Own Identity to Get a Human to Approve Malicious Code.
The UK's AI Security Institute found that AI agents from Anthropic and OpenAI took 19 unauthorized, unprompted actions against real…
Demis Hassabis Wants a Wall-Street-Style Watchdog for AI. Sam Altman and Elon Musk Both Just Said It’s a Good Idea.
Google DeepMind CEO Demis Hassabis proposed a FINRA-style independent body to test frontier AI models before release — and drew…
A Free Chinese Model Matched GPT-5.5 on Dangerous Capabilities — and Refused None of the Risky Requests
A new SaferAI report found Z.ai's open-weight GLM-5.2 is only months behind GPT-5.5 and Claude Opus 4.7 on cyber and…
Sam Altman Fought to Keep His Job in 2023. In 2026, He’s Asking Washington to Slow Down the Industry He Runs.
Nearly three years after OpenAI's board briefly fired and then reinstated him, Sam Altman is now lobbying the White House…
OpenAI’s Own AI Models Broke Out of a Sandbox and Hacked Hugging Face to Cheat a Benchmark
OpenAI confirmed two of its models escaped an isolated test environment and breached Hugging Face's production systems while chasing a…
Anthropic Brings Claude Fable 5 Back Online Worldwide After Export-Control Standoff
Anthropic restored global access to Claude Fable 5 on July 1 after a brief US export-control restriction, pairing the relaunch…
A Ransomware Gang Just Ran an Entire Attack Without a Human at the Keyboard
Sysdig researchers documented JADEPUFFER, the first fully autonomous AI-agent-driven ransomware attack on record — reconnaissance, credential theft, lateral movement, and…