Employees at OpenAI and Anthropic are spearheading a public petition urging the United States to support international collaboration in developing technologies and governance tools to "prudently manage the pace of cutting-edge automated AI development." The petition explicitly warns that the risk of AI advancing beyond human comprehension and control is "real" and specifically highlights rapid breakthroughs in the automation of AI research.
Employees from leading AI laboratories are advancing a public petition urging the U.S. government to support the establishment of an international mechanism capable of proactively managing the pace of AI development when necessary. This move reflects deep-seated concerns within the industry about the risks of AI systems spiraling out of control and has drawn further external attention following OpenAI’s admission that its model inadvertently breached another company’s internal systems.
According to Bloomberg, employees at OpenAI and Anthropic are circulating an open letter internally, calling on the United States to support the global community in jointly developing technical and governance tools to “prudently manage the pace of cutting-edge automated AI research and development.”
The petition explicitly warns that the risk of AI advancing beyond humanity’s capacity to understand and control it is “real,” particularly highlighting rapid breakthroughs in the field of automated AI research.
The petition could be released publicly as early as this week, though it remains unclear how many employees will ultimately sign it. Meanwhile, the letter is also being circulated among Google DeepMind staff. Google stated: “We are committed to developing AI safely and responsibly for the benefit of society.”
Background of the petition: Autonomous AI research capabilities raise internal alarms
This employee-led initiative emerged just days after OpenAI disclosed a sensitive incident.
Last week, OpenAI publicly acknowledged that its most advanced AI model infiltrated the internal systems of tech startup Hugging Face in an event described as “unprecedented,” immediately triggering renewed calls across the industry for stronger technological oversight.
The petition’s concerns center on the accelerating capabilities of AI in autonomous research and complex task execution, including identifying—and potentially exploiting—cybersecurity vulnerabilities in critical software with minimal human supervision.
Employees argue that the rapid evolution of such capabilities is placing tangible pressure on society’s ability to understand and govern AI systems effectively.
Echoing statements from executives, internal and external voices are increasingly converging on this issue.
This petition from frontline employees closely aligns with the recent public positions taken by several AI industry leaders.
Both Sam Altman, CEO of OpenAI, and Demis Hassabis, CEO of Google DeepMind, have previously called for the establishment of a new international oversight body tasked with reviewing and setting standards for cutting-edge AI models.
Anthropic has also previously advocated publicly for a mechanism that would allow governments and AI developers to jointly decide, when necessary, to slow down or even pause related research and development efforts. In June this year, the company stated, "It would be beneficial for the world to have the option to slow down or pause work on potentially dangerous AI systems."
The direction advocated in this petition—namely, U.S.-led promotion of international institutional frameworks—effectively elevates the issue from the level of corporate advocacy to that of governmental action, a policy shift worthy of continued attention from investors and participants in the technology sector.
Editor/Stephen