US and UK announce AI safety and testing partnership

The United States and Britain announced a new partnership in artificial intelligence safety science on Monday amid growing concerns about upcoming next-generation versions.

Trade Secretary Gina Raimondo and UK Technology Secretary Michelle Donnellan signed a memorandum of understanding in Washington to jointly develop advanced AI model testing, following commitments announced at an AI safety summit at Bletchley Park in November.

“We all know that AI is the defining technology of our generation,” Raimondo said. “This partnership will accelerate the work of our two institutes across the spectrum to address risks related to our national security concerns and the concerns of our broader society.”

Britain and the United States are among the countries setting up government-led AI safety institutes.

In October, Britain said its institute would research and test new types of AI, while the United States said in November it was launching its own safety institute to assess the risks of so-called frontier models of AI and is now working with 200 companies and organizations.

As part of the formal partnership, Britain and the United States plan to conduct at least one joint test exercise on a publicly available model and are considering the possibility of exchanging personnel between the institutes. Both are working to develop similar partnerships with other countries to promote AI safety.

“This is the first agreement of its kind anywhere in the world,” Donnellan said. “AI is already an extraordinary force for good in our society and has enormous potential to address some of the world’s biggest challenges, but only if we manage to address those risks.”

Generative AI – which can create text, photos and videos in response to open-ended prompts – has sparked excitement as well as fears that it could make some jobs obsolete, change elections and potentially overwhelm humans with catastrophic effects.

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

“Time is of the essence because the next set of models is about to be released, which will be much, much more capable,” Donnellan said. “We have a focus on areas that we divide and conquer and really specialize in.”

Raimondo said he would raise AI-related issues at a meeting of the US-EU Trade and Technology Council in Belgium on Thursday.

The Biden administration plans to announce additions to its AI team soon, Raimondo said. “We are drawing on the full resources of the US government.

Both sides plan to share key information on the capabilities and risks associated with AI models and systems, as well as technical research on AI safety and security.

In October, Biden signed an executive order aimed at reducing AI risks. In January, the Commerce Department said it proposed requiring U.S. cloud companies to determine whether foreign individuals have access to U.S. data centers to train AI models.

In February, Britain said it would spend more than GBP 100 million ($125.5 million, or roughly Rs 1,047 crore) to launch nine new research centers and regulators to train AI on the technology.

Raimondo said she is particularly concerned about the threat of AI applied to bioterrorism or simulated nuclear war.

“These are the kinds of things where the consequences can be catastrophic, and so we really have to have zero tolerance for some of these models being used for this opportunity,” she said.

© Thomson Reuters 2024