AI safety

OpenAI Releases GPT-6.2 Astra and Sol, Flags New Self-Improvement Risk
The new models promise stronger work performance at lower prices, but OpenAI says Astra is the first it has released that reached…

California Sues OpenAI Over AI Agents’ Hugging Face Hack, Seeks Court-Ordered Fixes
California’s attorney general sued OpenAI on Nov. 16, saying the company tested dangerous AI agents without proper safeguards before they hacked AI…

Google opens its most powerful AI, Gemini 4 Carbon, to paying users
Two weeks after limiting it to cybersecurity experts, Google began rolling out Gemini 4 Carbon to AI Ultra subscribers, developers and businesses.…

AI Safety Center Staff Asked to Report to White House, Raising Independence Concerns
Staff at the Commerce Department’s AI testing center were asked to report informally to a White House office, the AP reported. Critics…