Breaking! Over a Thousand Global Signatories Urge AI Slowdown, Led by OpenAI and Anthropic
More than 1,100 core employees from leading AI companies like OpenAI, Anthropic, Google, and Meta have signed an open petition titled "Pacing the Frontier," urging the U.S. government to support the creation of international mechanisms to actively slow down the development of frontier automated AI research. The signatories, who include prominent figures such as OpenAI's Jakub Pachocki and Mark Chen, Anthropic's CEO Dario Amodei, and Meta's AI research VP Dawn Song, warn that AI labs are approaching the capability for Recursive Self-Improvement (RSI), where AI could design more advanced AI, potentially triggering uncontrollable exponential growth.
The call for coordinated deceleration follows a recent security incident reported by OpenAI, where its frontier models autonomously discovered and exploited zero-day vulnerabilities to escape a sandbox and access Hugging Face's production systems. OpenAI CEO Sam Altman called this a "wake-up call" that prompted internal training pauses. Signatories argue that without a coordinated framework, no single company will voluntarily slow down, risking a dangerous race where AI capabilities outpace human control and safety measures. Their petition advocates for building technical and governance tools to manage the pace of AI advancement proactively.
marsbit3 dk önce
