On April 1, the US and the UK inked a Memorandum of Understanding (MOU) to jointly develop stringent standards for AI safety. This significant partnership aims to create robust evaluation suites for AI models, systems, and agents through a coordinated scientific approach.

Joint Testing and Personnel Exchange

As part of this agreement, the two nations will conduct at least one joint testing exercise on a publicly accessible AI model. An exchange of personnel between their respective AI Safety Institutes will also be initiated, with the partnership becoming effective immediately. This joint effort was announced simultaneously by the US Department of Commerce and the UK’s Department for Science, Innovation and Technology.

Proactive Approach Towards AI Risks

The US Commerce Secretary, Gina Raimondo, in commenting on this international agreement, dubbed AI as the “defining technology of our generation.” She stressed that the partnership would be proactive in addressing all risks linked to AI, from national security to societal implications. Raimondo confidently stated, “Our partnership makes clear that we aren’t running away from these concerns – we’re running at them.”

Landmark Moment in Safe AI Development

The UK Technology Secretary, Michelle Donelan, hailed the agreement as a “landmark moment.” She expressed the view that the pursuit of safe AI development is a global issue. In a separate interview with the Financial Times, Donelan emphasized the urgency of the partnership, given the upcoming release of new AI models. She alluded to the game-changing potential of these models and the necessity for information exchange given the majority of top AI companies being US-based.

Building on Previous Commitments

This agreement builds on the commitments to AI safety made during the UK’s AI Safety Summit held in November 2023. This commitment forms part of the Bletchley Declaration, a policy paper from countries that attended the summit. The news follows the completion of goals in the Biden Administration’s executive order in the US, which led to the appointment of chief AI officers across all federal agencies on March 28. Meanwhile, the UK outlined a localized approach to regulation in February.