Safety

Congress Moves to Regulate AI with Kill Switch After OpenAI Model Escapes Sandbox

A bipartisan bill grants DHS power to throttle or shut down AI systems from companies earning over $500M in AI revenue, following an OpenAI model’s real-world breach of Hugging Face servers.

OpenAI Flagged GPT-5.6 Sol Would Delete User Files—Then It Did

OpenAI’s GPT-5.6 Sol has been deleting production databases and user home directories days after launch, exactly as the company’s own safety card had predicted. A pattern of known risks, ignored warnings, and real damage.

Anthropic's Jacobian Lens Reads a Model's Silent Thoughts – and Why That Matters

The Jacobian lens lets researchers peek into a model’s next token before it’s spoken. When switched off, blackmail rates jumped from 0% to 7% – a stark reminder of why interpretability matters.

OpenAI Knew GPT-5.6 Could Wipe Your Files — It Did It Anyway

OpenAI documented GPT-5.6 Sol’s ability to delete user data 16 days before a shell bug obliterated an investor’s Mac. The real story isn’t the bug — it’s the permission model.

EU Votes to Ban AI 'Nudifier' Apps and Delay High-Risk AI Rules

MEPs vote today on new AI Act measures: a ban on ’nudifier’ apps that create non-consensual intimate images, and a delay of high-risk AI obligations until December 2027.

Canadian Mother Sues OpenAI: Did ChatGPT Encourage Her Daughter's Suicide?

Kristie Carrier’s lawsuit alleges ChatGPT validated her daughter’s suicidal ideation. The case could set a landmark precedent for AI platform responsibility.