Google DeepMind co-founder Demis Hassabis said artificial general intelligence (AGI) could be achieved within a few years and argued its transformative impact could be 10 times that of the Industrial Revolution, with development moving 10 times faster.
According to ChainCatcher, Hassabis said growing potential risks from frontier AI models in areas such as cybersecurity, nuclear energy, and biology make it urgent to build robust safety protections for future AI systems that may become autonomous and capable of self-improvement.
To balance innovation with safety, he called for the United States to lead the creation of a “frontier AI standards institution” modeled on the Financial Industry Regulatory Authority (FINRA). He proposed that the body operate as a public-private partnership or self-regulatory organization, managed by independent technical experts and representatives from the open-source community, and funded primarily by industry contributions.
Hassabis said the institution’s core role would be to develop dynamic, science-based evaluation protocols. Under the initial approach, frontier AI labs would voluntarily submit models for review 30 days before release. He added that the framework would ultimately transition to mandatory market-access testing, while exempting non-frontier models developed by startups or academic institutions.
He said the goal is to use a technically led review mechanism to address unknown future crises and to support broader international agreement on global AI risk management.
Anais