US and UK governments agree deal to collaborate on AI safety testing

The U.S. and U.K. governments have agreed to cooperate on artificial intelligence safety, signing a proper agreement that can see them collaborate on methods for testing AI models and assessing the risks they pose.

The agreement, signed today in Washington D.C. by U.K. Science Minister Michelle Donelan and U.S. Commerce Secretary Gina Raimondo, lays out how the partners will pool their technical expertise, information and talent on AI questions of safety.

In line with Reuters, the deal is the primary bilateral agreement on AI safety on this planet, and it comes at a time when various governments have been pushing hard to extend regulation within the AI industry, amid fears that the technology may very well be misused in various ways. A few of the risks include AI getting used to stage damaging cyberattacks or design recent bioweapons, while luminaries akin to Elon Musk have warned that the technology might “kill us all.”

The deal will enable the U.K.’s recent AI Safety Institute, which was established in November, and its U.S. equivalent, which has not yet been formed, to share their expertise and know-how via secondments of researchers from each countries. As well as, the institutes are also expected to work together to judge private AI models built by firms akin to OpenAI and Google LLC.

Donelan told the Financial Times that the partnership is modeled on an existing one between the U.K.’s Government Communications Headquarters or GCHQ and the U.S. National Security Agency, which regularly collaborate on matters referring to intelligence and national security.

“The subsequent yr is after we’ve really got to act quickly because the following generation of [AI] models are coming out, which may very well be complete game-changers, and we don’t know the total capabilities that they may offer yet,” Donelan told the Financial Times. “The incontrovertible fact that america, a fantastic AI powerhouse, is signing this agreement with us, the UK, speaks volumes for a way we’re leading the way in which on AI safety.”

Donelan said it’s necessary for the U.K. to partner with the American government, since lots of a very powerful AI firms are based within the U.S. So its expertise is vital to understanding the risks posed by AI. As well as, the U.S. government holds greater sway over those firms, so it’s higher capable of ensure those firms adhere to existing commitments they’ve made on AI safety.

Under Prime Minister Rishi Sunak, the U.K. has tried to take leading role in marshaling nations to give you an efficient international response to the growing power of AI models, that are sometimes known as “frontier AI.” It convened the world’s first ever AI Safety Summit and it has committed £100 million ($125 million) towards its recent AI Safety Institute – an amount that dwarfs the somewhat meager $10 million commitment by the U.S. to its own AI Safety Institute.

Strangely, though, Donelan told the Financial Times that the U.K. government doesn’t have any plans to manage the event of AI more broadly, not less than not within the near term.

The U.K.’s stance differs from many other nations and regions. The European Union last yr rolled out its AI Act, which is taken into account to be the hardest regulatory regime regarding AI development on this planet. U.S. President Joe Biden has issued an executive order targeting AI models which have the potential to threaten national security, ahead of further laws that’s still within the pipeline. China has also issued guidelines on AI that aim to be certain that the technology doesn’t sidestep its longstanding censorship policies.

The collaboration between the U.S. and U.K. is a welcome initiative, as safety around AI has generally taken a backseat until recently, said Andy Thurai, vice chairman and principal analyst at Constellation Research Inc. The EU’s AI Act was really the primary major initiative to attempt to rein in AI developers, he said, defining 4 categories of AI models based on the extent of risk they pose and outright banning certain functions, akin to facial recognition and biometrics, in some specific situations.

In line with Thurai, it will be a lot better for the U.S., U.K and EU to come back together and collaborate on guidelines, as AI poses a world risk, so today’s initiative is start.

“But the reality is that not enough is being done. Initiatives akin to C2PA, which goals to determine authenticity and provenance for AI-generated content, are worthwhile and I might like to see more,” Thurai said. “Unfortunately, few persons are willing to do the suitable thing and fund such projects to create truly responsible AI. As an alternative, we’re in an AI arms race and everybody involved is racing to attempt to construct larger, higher and faster models that outperform the competition. So it would be some time before we see AI safety gaining any real momentum.”

At the least, it’s likely that the security tests that can be developed by the U.S. and the U.K. will play a job in helping legislators and technology executives to mitigate the risks posed by fast-evolving AI systems, and a few of the leading players in AI appear willing to cooperate. Earlier this yr, Google, OpenAI and Anthropic PBC all published plans outlining how the security tests being designed will inform their future product development strategies.

Image: Freepik