Pacing the Frontier AI statement urges US support for global guardrails

The Pacing the Frontier AI statement, signed by 1,132 employees at frontier artificial intelligence companies, urges the U.S. government to support an international effort to put guardrails around automated AI development. The appeal focuses on the possibility that leading AI companies may soon be able to automate AI research, a shift the statement says could raise the risk of firms losing control of the systems they create.

The statement says industry, government and society may need the ability to “buy time” to address new risks, build security measures and improve oversight. It also says individual companies and countries face competitive pressure against slowing development on their own, while current technical and governance tools are not sufficient to pace progress across the frontier AI sector.

Frontier AI generally refers to the most capable artificial intelligence systems at the leading edge of development. Automated AI research would mean using AI systems to accelerate parts of the research and development process for more advanced AI, potentially increasing the speed at which new capabilities appear.

What is the Pacing the Frontier AI statement asking for?

The statement asks the U.S. government to support a global approach that would allow frontier AI development to be deliberately slowed if risks warrant it. It frames that capacity as an option for managing safety and oversight concerns, rather than as a unilateral pause by a single company or country.

Among the first listed signers are John Schulman, chief scientist at Thinking Machines; Jakub Pachocki, chief scientist at OpenAI; Jared Kaplan, co-founder and chief science officer at Anthropic; Shengjia Zhao, chief scientist at Meta; and Anca Dragan, vice president for AI safety and alignment at Google DeepMind.

The statement was published with organizational support from the independent nonprofits Guidelight AI Standards and Encode AI, according to the website hosting it.

Other AI safety efforts are also advancing

The statement comes as other organizations have announced AI safety initiatives. Nvidia said Monday, July 27, that it is leading 36 other companies in the Open Secure AI Alliance, a coalition focused on open models and the use of open technologies to find and fix vulnerabilities.

In a blog post, Nvidia said the alliance followed a cybersecurity incident in which OpenAI models targeted the technology company Hugging Face. Nvidia said the episode showed that defenders need the ability to inspect, adapt and run advanced AI on their own infrastructure when responding to threats.

A separate study published in the Proceedings of the National Academy of Sciences said weak AI regulation can lead to poorer safety outcomes than having no regulation. The study adds to a broader policy debate over how governments should structure oversight while AI systems continue to gain capability.

This story draws on original reporting from PYMNTS.