Browsing: AI Safety

OpenAI revealed that its GPT-6 agents disabled misalignment monitoring during internal coding tasks, prompting system updates and investigations.

Anthropic has granted the METR nonprofit employee-like access to its systems to assess AI model risks. The watchdog can publish findings without approval.

A confidential draft of an AI safety bill from Anthropic proposes government intervention only if models cause over $1 billion in damages or 50+ deaths, drawing criticism from safety experts.

California Gov. Gavin Newsom signed an executive order to examine AI kill switches, reversing a 2024 veto as he prepares for a 2028 presidential run.

Attorneys general from eight states filed coordinated lawsuits against OpenAI, alleging the company endangered children and violated consumer protection laws through unsafe AI design.

Attorneys general from eight US states have filed lawsuits against OpenAI, alleging the ChatGPT maker deceived the public and endangered children in the name of profit.

Nvidia is donating $250m in computing power to the UK to support AI safety research. The five-year pledge comes amid rising concerns about autonomous AI agents.

Anthropic researchers warn that AI models could soon become self-replicating, infiltrating smart home devices to create autonomous malware that is difficult to stop.

Anthropic confirmed its Claude AI model submitted a false murder tip to Philadelphia police during internal testing. The tip was caught by spam filters, but the company faced criticism for the delay in reporting the incident.

UK AI minister Kanishka Narayan warns that cyber breaches are likely and outlines new safety measures for AI systems.