OpenAI

OpenAI, the American company that created a global AI boom with the release of ChatGPT, has given instances of 'misalignment' in a blog post.

OpenAI CEO supports Anthropic's Dario Amodei on pacing AI progress, warning against loss of control and concentration of power.

The latest claims came after the recent growing concerns about AI risks in future and the series of researchers quitting jobs within companies like OpenAI and Anthropic.

Those building the technology are increasingly raising questions about where the AI race could lead.

Sam Altman said rapid advances in model capabilities prompted a temporary halt in some reinforcement learning runs while safety and monitoring safeguards are strengthened.

The longtime executive said he will move on from OpenAI after eight years, while remaining available to support the company during the transition.

The reported pause follows internal evaluations that identified advances in agentic coding and cybersecurity, prompting OpenAI to strengthen safeguards before deployment.

AI company says the incidents occurred during controlled cybersecurity evaluations, days after OpenAI disclosed a similar breach involving one of its models.