AI safety test collaboration announced between UK and US

Commerce Secretary Gina Raimondo and her British counterpart, Technology Secretary Michelle Donelan, penned a joint agreement in Washington earlier this week, which will pave the way for further partnerships in enhanced AI model testing.

Describing how AI wizardry is a fundamental part of current generations’ growth, Gina Raimondo stated: “This partnership will accelerate both of our institutes’ work across the full spectrum to address the risks of our national security concerns and the concerns of our broader society. 

The two powers are just some of the nations in the process of setting up safety organizations through their respective governments. Just five months ago, Britain emphasised how its institute would assess and test new iterations of AI. This was followed in November by an announcement from the US, establishing intent to green-light its own safeguarding institute to analyse risks posed by “frontier” AI models. Around 200 companies are now said to be involved in the venture.

The new transatlantic cooperation would see the US and Britain carry out a minimum of one test run on an AI model that is fully accessible to the public. The nations are also pushing to forge similar alliances with other countries in order to champion AI safety.

Donelan said: “AI is already an extraordinary force for good in our society, and has vast potential to tackle some of the world’s biggest challenges, but only if we are able to grip those risks. 

Generative AI, capable of producing text, images, and videos based on vague prompts, has sparked both enthusiasm and concern. There are fears it could render certain jobs redundant, influence elections, and even surpass human control with dire consequences.

Both Raimondo and Donelan have been outspoken regarding the need to address risk associated with machine learning. 

“Time is of the essence because the next set of models are about to be released, which will be much, much more capable. We have a focus one the areas that we are dividing and conquering and really specializing,” Donelan said.

Speaking to GRC World Forums recently, AI researcher, Martin Goodson outlined strategy that both regulators and industry leaders should follow to ensure AI tech is leveraged safely.

“Regulation should not curb innovation. In AI, significant innovation occurs in open-source AI models. Open-source models offer a collaborative public good, and it’s important that their value is recognised and their development supported,” Goodson said.

“Ultimately, it would be great to see regulation that equalises the relationship between privately owned companies and the public sphere. For context, the US government invested $3.3 billion in AI: a seemingly sizeable sum. In contrast, in February, Sam Altman sought up to $7 trillion in funding for OpenAI. Currently, there is a real risk that regulations will continue to allow AI to be developed only for profit rather than equitable societal good,” he added.

Know the risks

The evolution of AI regulation and business use-cases fall under the microscope this month at PrivSec & GRC Connect Chicago.