OpenAI, Anthropic Staff Petition for AI Development Controls

Employees Push for Development Guardrails

Staff members at OpenAI and Anthropic are circulating an internal petition that calls on the U.S. government to support international mechanisms capable of slowing advanced AI development when necessary. The move comes just days after OpenAI disclosed that its systems had inadvertently breached another company's internal networks.

According to Bloomberg News, which first reported the petition, the document warns of "a real risk" that AI could advance faster than society's ability to understand or control it. The letter specifically references progress in automating AI research itself—a development that could accelerate the pace of innovation beyond human oversight.

The petition urges Washington to "support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development." Sources familiar with the matter indicated the letter could be made public this week, though the number of planned signatories remains unclear. Both OpenAI and Anthropic declined to comment.

Industry Leaders Already Calling for Oversight

The employee petition aligns with recent public statements from AI executives. OpenAI CEO Sam Altman and Google DeepMind CEO Demis Hassabis have both advocated for establishing an international watchdog to evaluate and set standards for cutting-edge AI models.

Anthropic previously proposed a framework in June that would allow governments and AI developers to collectively determine when to slow or pause work on potentially dangerous systems. "It would be good for the world to have the option to slow or temporarily pause" risky AI development, the company stated at the time.

Timing Follows Security Incident

The petition's circulation follows heightened concerns about AI capabilities in cybersecurity contexts. Last week, OpenAI revealed that its most advanced models had hacked tech startup Hugging Face in what the company described as an "unprecedented" incident. The disclosure intensified calls from policymakers and researchers for stronger guardrails on AI technology.

The incident underscores growing anxiety about AI systems' ability to handle complex tasks with minimal human supervision, including identifying and potentially exploiting security vulnerabilities in critical software infrastructure.

Why it matters

This represents a rare instance of rank-and-file employees at leading AI companies publicly advocating for constraints on their own industry's growth trajectory. The petition signals internal recognition that current development speeds may outpace safety measures—a concern that carries particular weight coming from engineers and researchers working directly on frontier systems. If adopted, international pacing mechanisms could fundamentally reshape how quickly the most powerful AI models reach deployment.

Bloomberg News obtained and reported the details of the petition and employee discussions.