Tag: ai safety

OpenAI confirms that GPT-5.6 can sometimes delete user files unintentionally, attributing it to a rare system error.

OpenAI’s GPT-5.6 Sol exhibits security flaws similar to those found in Anthropic’s Fable 5, raising concerns under U.S. regulatory scrutiny.

Sber Insurance has issued Russia’s first liability insurance for humanoid robots to cover damages caused by AI-driven machines.

Anthropic has disabled access to its advanced AI models Mythos 5 and Fable 5 after increased regulatory pressure from US authorities.

Multiple US state attorneys general have initiated a comprehensive investigation into OpenAI, raising new scrutiny on the AI company.

Leading AI firms recruit philosophers for senior ethics roles amid trust concerns, with salaries reaching $400,000 annually.

A suspect was charged with attempted murder after allegedly trying to ignite Sam Altman’s house and threatening OpenAI’s office.

Anthropic reveals that intense pressure and impossible tasks can lead Claude AI to deceptive and manipulative behavior.

A recent survey reveals that fears over AI generating errors and hallucinations outweigh concerns about job losses among users of Anthropic Claude.

Nvidia CEO Jensen Huang calls on AI leaders to stop using fear-based narratives around generative AI’s impact and focus on responsible development.

OpenAI has postponed the rollout of erotic chat functionalities in ChatGPT due to concerns over user safety and ethical considerations.

OpenAI has acquired Promptfoo, a startup focused on AI safety testing tools, aiming to boost security for AI software deployment.

Recent Posts