The U.S. and the UK signed a memorandum of understanding (MOU) Monday stating that the two will collaborate to test the safety of artificial intelligence systems.
The agreement follows promises made at the UK’s AI Safety Summit last year, where the UK and the U.S. pledged that their respective AI Safety Institutes would formally partner on AI security.
Michelle Donelan, the UK’s secretary of state for science, innovation and technology, said the partnership will set a precedent for other nations looking to understand and regulate the technology.
“We have always been clear that ensuring the safe development of AI is a shared global issue. Only by working together can we address the technology’s risks head on and harness its enormous potential to help us all live easier and healthier lives,” Donelan said in a statement. “The work of our two nations in driving forward AI safety will strengthen the foundations we laid at Bletchley Park in November, and I have no doubt that our shared expertise will continue to pave the way for countries tapping into AI’s enormous benefits safely and responsibly.”
Donelan and Gina Raimondo, the U.S.’s secretary of commerce, signed the agreement in Washington. In her remarks, Raimondo said the partnership should prove to be a win-win for both parties.
Related Stories
“AI is the defining technology of our generation. This partnership is going to accelerate both of our institutes’ work across the full spectrum to address risks within both our national security and broader society,” she said Monday. “Our partnership makes clear that we aren’t running away from these concerns—we’re running at them. Because of our collaboration, our institutes will gain a better understanding of AI systems, develop, and conduct more robust evaluations and provide more rigorous and useful guidance.”
Raimondo noted that signing the MOU is “in line with the Biden administration’s approach to AI: managing its risks so that we can harness its benefits.” To date, the U.S. has not passed national legislation that would regulate AI’s impact on business, though President Biden signed an Executive Order in October as a starting point for technology developers. To date, much of the U.S.’s actions around AI have been concerned with national security. It seems this partnership falls into that bucket, as well.
The UK has adopted a cross-industry regulatory framework on AI. Meanwhile, the EU recently adopted the landmark EU AI Act, which experts expect could influence future legislation in both the UK and the U.S.
For now, as the UK and U.S. develop their respective strategies, the institutes have plans to develop a shared approach to AI system evaluation, collaborate on technical research around AI safety, open the door for policy alignment, share information and develop similar partnerships with other interested countries.
If the November conference in the UK is any indication of which countries may be interested, partnerships with Australia, Canada, Ireland, Singapore, Turkey and the United Arab Emirates could soon be pending. Those countries—as well as a slew of others—signed what’s known as the Bletchley Declaration in November. The signatories committed to a global dialogue around safety and policy for AI.
Per Raimondo and Donelan, the additional support would be welcome.
“We anticipate that in the weeks and months ahead, more partnerships will help create a global network of AI safety, built through numerous linkages between government-backed scientific institutions that are doing research on and developing standards and guidelines for the safe development and implementation of advanced AI systems,” Raimondo noted in her remarks. “That will include our safety institutes but may also incorporate other types of scientific organizations. We want to think outside the box. We want to think big.”