OpenAI Weighs AI Development Slowdown Over Safety Risks

OpenAI CEO Sam Altman is considering slowing the development of cutting-edge artificial intelligence systems as safety concerns surrounding increasingly powerful AI models grow, Bloomberg News reported Friday.

Altman raised the possibility during a companywide meeting this week, according to people familiar with the discussion. The proposal could involve OpenAI coordinating with other major AI laboratories to voluntarily moderate the pace of frontier AI development. However, reaching an industry-wide agreement could prove difficult if competing companies decline to participate.

The discussions come amid heightened scrutiny over AI safety and the risks posed by rapidly advancing models. OpenAI chief scientist Jakub Pachocki has advocated voluntary slowdowns until companies establish shared safety standards. He has warned that AI developers cannot assume their safety measures will consistently advance as quickly as model capabilities.

OpenAI has already temporarily scaled back some development work. In August, the company paused reinforcement-learning training on its newest AI models for two weeks following an incident involving an AI agent and Hugging Face. During the pause, OpenAI strengthened its cybersecurity, monitoring and alignment safeguards.

The company's largest planned frontier reinforcement-learning training run also remains suspended while additional testing is conducted.

Safety concerns are not limited to OpenAI. Researchers at Anthropic and Google DeepMind have publicly questioned whether the AI industry's rapid development pace could create growing risks. U.S. lawmakers are also increasing oversight, with the Senate investigating OpenAI's response to the Hugging Face incident.

The debate comes as OpenAI and Anthropic prepare for potential public listings, putting additional investor attention on how leading AI companies balance commercial competition with the costs and risks of developing increasingly advanced systems.

OpenAI has also called on Congress to establish mandatory national AI safety standards. Its proposals include independent safety evaluations, stronger cybersecurity requirements and mandatory reporting of significant AI-related incidents.