Google DeepMind CEO Demis Hassabis is urging the US government to create a federal standards organization that would review advanced AI models for national security risks, including cybersecurity vulnerabilities and biological threats. The Nobel laureate published his proposal on X Tuesday, warning that artificial general intelligence—systems that match or exceed human cognitive ability—requires urgent oversight before risks escalate.

Hassabis envisions a public-private partnership modeled on the Financial Industry Regulatory Authority, with a board of independent technical experts and open-source representatives. Under his plan, frontier AI labs would initially submit models voluntarily for review up to 30 days before release, with mandatory compliance for US market deployment once the body proves effective. Testing would target safety guardrail bypass attempts, deceptive behavior, and enforce practices like digital watermarking of AI-generated content. The organization would need substantial industry funding to attract top technical talent and acquire computing resources for large-scale evaluations.

The call comes amid intensifying US-China competition in AI development. Chinese models from DeepSeek and Z.ai are gaining traction among American companies as costs rise, prompting US lawmakers to consider curbing adoption of foreign systems. The State Department has described Chinese AI proliferation as a serious concern. Hassabis argues the US is uniquely positioned to lead on AI governance given its economic and technical stature, and his proposal follows similar appeals from Anthropic CEO Dario Amodei and OpenAI's Sam Altman for coordinated international standards.

Source: CNBC