The head of Google's artificial intelligence unit urged the United States to take the helm of a new standards organization that would supervise emerging AI models and evaluate potential dangers to national security such as cyber attacks and biological threats.
Demis Hassabis, a Nobel laureate and Google DeepMind chief, wrote in a post on X on Tuesday that "urgent action" was necessary and emphasized the necessity of tackling the dangers linked to artificial general intelligence - the stage where AI equals or exceeds human cognitive abilities.
"We've already seen the challenges frontier models pose for cybersecurity, and other threats including nuclear and bio risks may soon emerge as capabilities continue to advance," he said.
Hassabis suggested forming a partnership between the public and private sectors, led by the U.S., under federal oversight. Officials at the White House, State Department, and Commerce Department were contacted for comment.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
About a month ago, CNBC reported that Hassabis and Anthropic CEO Dario Amodei jointly urged the creation of a U.S.-led coalition to develop AI regulations and standards during a G7 gathering that featured heads of state and tech executives, including President Donald Trump. Sam Altman of OpenAI likewise proposed an analogous organization in a Financial Times piece earlier this month.
Hassabis stated that the United States is ideally placed to take the lead on developing an AI framework "given its economic and technical standing."
"It could establish a new Standards Body modelled on a federally overseen public-private partnership or self-regulatory organisation, much like the Financial Industry Regulatory Authority (FINRA), with a board that includes independent leading technical experts and open-source representatives," he added.
FINRA supervises brokerages and securities exchanges across the country. The proposed body would need "substantial" funding "in order to attract world-class technical talent and provide the necessary compute resources for large-scale testing," Hassabis said. He added that the necessary financing would probably be provided by the private sector.
Under the plan, leading AI labs would initially submit their models for a voluntary review period of up to 30 days prior to public release; once the process is deemed "effective," such submission would become a required step for any U.S. market deployment.
"Specific agentic AI tests could look for attempts to bypass safety guardrails or signs of deception, and ensure best practices, such as digitally watermarking AI-generated images and generating human-readable output tokens to understand model reasoning," Hassabis said.
The push for stronger regulation comes amid an intensifying competition between the United States and China in building and rolling out AI systems. Chinese companies like DeepSeek and Z.ai have released AI models that many view as rivaling the top-tier offerings from Anthropic and OpenAI. These systems are increasingly being adopted by American businesses, partly because of the rising cost of AI development.
As a result, U.S. lawmakers are exploring measures to limit the increasing use of Chinese AI systems within American companies - a development the State Department called "serious concerns" in a CNBC statement.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets
