US and UK announce partnership on AI safety and testing

The United States and the United Kingdom on Monday announced a new partnership on artificial intelligence safety science as concerns grow over upcoming next-generation versions.

Following the commitment announced at the AI ​​Security Summit in Bletchley Park in November, Business Secretary Gina Raimondo and UK Technology Secretary Michelle Donelan signed a memorandum of understanding in Washington to jointly develop Advanced AI model testing.

“We all know that artificial intelligence is the defining technology of our generation,” Raimondo said. “This partnership will accelerate the full range of work across our two agencies to address risks to our national security and broader societal issues.”

The UK and the US are among the countries establishing government-led AI security agencies.

The UK said in October that its institute would examine and test new types of AI, while the US said in November it was setting up its own security institute to assess the risks of so-called cutting-edge AI models and is currently working with 200 companies and entities cooperate.

Under the formal partnership, the UK and US plan to conduct at least one joint testing exercise on publicly accessible models and are considering exploring personnel exchanges between agencies. Both countries are committed to building similar partnerships with other countries to promote AI safety.

“This is the first agreement of its kind in the world,” Donelan said. “Artificial intelligence is already an extraordinary force for good in our society and has enormous potential to solve some of the world’s biggest challenges, but only if we can address these risks.”

Generative AI, which can create text, photos and videos based on open-ended prompts, is generating both excitement and concern that it could make some jobs obsolete, upend elections, and potentially overwhelm humans with catastrophic effects.

In a joint interview with Reuters on Monday, Raimondo and Donelan said urgent joint action was needed to address AI risks.

“Time is of the essence because the next set of models are about to be released and they’re going to be much more powerful,” Donelan said. “Our focus is on areas where we’re dividing our labor and conquering and really specializing.”

Raimondo said she would raise the issue of artificial intelligence at a meeting of the U.S.-EU Trade and Technology Council in Belgium on Thursday.

Raimondo said the Biden administration plans to announce new additions to its artificial intelligence team soon. “We are mobilizing the full resources of the U.S. government.”

The two countries plan to share key information on capabilities and risks associated with AI models and systems, as well as research on AI safety and security technologies.

In October, Biden signed an executive order aimed at reducing the risks of artificial intelligence. In January, the Commerce Department said it proposed requiring U.S. cloud companies to determine whether foreign entities are accessing U.S. data centers to train artificial intelligence models.

The UK said in February this year that it would spend more than 100 million pounds ($125.5 million) to establish nine new research centers and train regulators on artificial intelligence technology.

Raimondo said she is particularly concerned about the threat of AI being used in bioterrorism or nuclear war simulations.

“The consequences of these things can be catastrophic, so we really have to have a zero-tolerance approach to certain models being used for this kind of functionality,” she said.