• by Laurie Sullivan , Yesterday

Demis Hassabis, Nobel Laureate and Google DeepMind co-founder and CEO, called for "urgent action" and oversight in the form of a framework to address risks that may surface as society moves further into the use of artificial general intelligence (AGI), AI with human-level cognitive abilities.

In a blog post, Hassabis laid out his vision for a U.S.-led AI initiative requiring a new approach to economic and technical standards that would test capabilities for frontier AI models, which describes the most advanced, high-performing, and general-purpose artificial intelligence (AI) systems available at any given moment.

Hassabis' post comes one day after Microsoft CEO Satya Nadella's post on how AI creates new risks for companies with regard to consumer data privacy and intellectual property.

Nadella’s argument -- "The Reverse Information Paradox," published in a post on X -- focuses on the issue of stealing trade secrets and how AI requires more data protection -- more than patent and intellectual property safeguards.

Hassabis describes the first step in developing this framework as creating a new Standards Body, modeled on a federally overseen public-private partnership or self-regulatory organization similar to the Financial Industry Regulatory Authority (FINRA).

Its board would include leading independent 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,” he wrote.

The model would qualify as “Frontier-class” if it meets certain criteria based on benchmarks determined by the Standards Body that updates regularly keep pace with change.

Participating organizations with Frontier Models, defined by benchmarks, would be deemed “Frontier Labs” and would 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.

The industry would set guidelines for the framework. For example, Frontier Labs would voluntarily share models with the Standards Body for review up to 30 days before release.

Once the assessment protocol is shown to be effective, the standards would then be run in the U.S. market.

Labs would also work with the Standards Body to address any critical post-release vulnerabilities, similar to the existing standards by the Interactive Advertising Bureau in partnership with the IAB Tech Lab, which supports the advertising industry.

“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,” he wrote.

This framework would create a completely new cottage industry. He described how the body would work with the U.S government by promoting an ecosystem of third-party auditors to help with assessments and development of new benchmarks and evaluations.

"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," Hassabis wrote. "On the horizon, we will need robust safeguards to maintain control of increasingly agentic, recursively self-improving systems - and tackle unknown issues that will only become clearer over time."

Hassabis did not provide details in his essay, which was written and published Tuesday in a blog post, "A Framework for Frontier AI and the Dawning of a New Age," about how his framework will influence or affect the advertising industry, but rather focuses on corporate ad mechanics. It outlines that AGI requires completely rewriting the rules of traditional business and economics.

Google's executive leadership, including Philipp Schindler, Google chief business officer, in the past has made it clear that any breakthrough from DeepMind is immediately put on a "direct pipeline from labs to business results" and become core to the company’s operating strategy.

The Standards Body would take responsibility to develop assessment protocols and work with appropriate federal agencies and the US National Labs to conduct testing in areas relevant to national security.

"The magnitude of this technology’s impact will be unprecedented, perhaps 10x of the Industrial Revolution at 10x the speed," he wrote in the post. "It will help us solve some of the biggest problems society faces from accelerating drug discovery to developing new clean energy sources to creating novel advanced materials. We could even reach a point where resources are no longer the limiting factor for human progress, leading to an amazing new era of abundance."

The world evolves around a “large degree of uncertainty and the stakes are this high,” so moving forward with “cautious optimism is the sensible and correct strategy” is called for during this time.

“That calls for public policy that promotes innovation while also incentivizing responsibility and security, fosters international collaboration on key safety issues, and encourages careful consideration of how AI is deployed for the benefit of society,” he wrote.