Google DeepMind CEO: AGI to Arrive Within Years, Sparking Changes as Great as Electricity and Fire's Discovery

Demis Hassabis, CEO of Google DeepMind, argues that AGI could emerge within years, urging the establishment of an independent standards body to evaluate advanced AI models before their release. He compares AGI's potential impact to revolutionary discoveries like electricity and fire, emphasizing the need for safety measures that match the pace of technological advancement. Hassabis warns that even current AI models pose cybersecurity threats, with risks expanding into biology, nuclear technology, and national security. He stresses the necessity of control mechanisms for self-improving AI systems and proposes a U.S.-based 'Frontier AI Standards Body' modeled after FINRA, combining industry and expert collaboration to assess AI risks before deployment. Hassabis advocates for mandatory safety verification for the strongest AI systems, highlighting the need for dynamic yet rigorous testing frameworks to guide global regulatory standards.