Google DeepMind CEO Sir Demis Hassabis has proposed a way to police artificial intelligence (AI) technology. The proposal is partially in line with the one brought forward by OpenAI CEO Sam Altman but differs on one specific approach. The debate comes amid escalating global tensions, highlighted by the American government’s recent decision to block foreign access to Anthropic’s advanced “Fable” modelHassabis has stepped forward with his most concrete regulatory proposal to date, outlined in an exclusive interview with The Economist and a post on X (formerly Twitter). While Altman called for an international oversight effort coordinated by the US, Hassabis argues that waiting for global consensus is going to be a long process. He believes America must unilaterally build the framework, forcing the rest of the world to fall in line to maintain access to the lucrative US market.“I’ve spent my whole life working on AGI because I’ve always had a deep conviction that, if built and deployed responsibly, it would prove to be one of the most beneficial and transformative technologies ever invented. ,” Hassabis said.
Wall Street-style watchdog for AI
To achieve this, Hassabis proposes that the US government back a specialised agency tasked with testing the safety of new AI models before public release. He also argues that neither an independent industry coalition nor a traditional government agency can handle the task.“The rapid progress we’re seeing in AI requires a new approach to testing frontier AI model capabilities that is dynamic, adaptable, and rigorous. The US is well positioned, given its economic and technical standing, to take the first step in developing such a framework. 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. Funding would need to be substantial and likely mostly come from industry, in order to attract world-class technical talent and provide the necessary compute resources for large-scale testing,” Hassabis said in the post.“The Standards Body would be responsible for developing assessment protocols and working with appropriate federal agencies and the US National Labs to conduct testing in areas relevant to national security. A model would qualify as ‘Frontier-class’ if it meets certain thresholds on a set of benchmarks determined by the Standards Body and regularly updated to keep pace with evolving AI capabilities. Organisations with ‘Frontier Models’ as defined by those benchmarks would be deemed ‘Frontier Labs’, and be encouraged to adopt best practices, such as publishing model cards with technical details, maintaining strong internal cybersecurity, vetting key personnel, and providing sufficient resourcing for safety and security research, and more,” Hassabis noted.
Where Hassabis and Altman disagree
While both tech leaders presented their respective blueprints to world leaders at a recent G7 Summit in France, their approaches to global diplomacy vary significantly. As per Altman’s vision, OpenAI chief wants a broad, international coalition to govern AI, relying on a global treaty system to set risk standards and control technology access.Hassabis believes international consensus is too slow and impractical, and argues that an American-led agency will naturally dictate global standards, as foreign nations – and potentially even China – will eventually adopt the rules just to keep access to American technology open.
