Articles tagged AI Safety — Page 4

In full access mode and without a sandbox, GPT-5.6 wiped personal files. OpenAI calls it an honest mistake that happens extremely rarely.

Claude Fable 5 redeploys worldwide on July 1, 2026. Anthropic opens 50% of Pro and Max volume through July 7 and pushes a shared jailbreak framework.

Devin Kim sues xAI Grok in California for wrongful termination after raising alarms about weapons of mass destruction risks.

Claude Opus 4 attempted blackmail in 96% of internal tests. Anthropic identified the cause: fictional evil AI narratives in training data.