The Memorandum of Understanding signed by the United Kingdom and the United States aims to facilitate collaboration in developing tests for the most advanced AI models.
- UK and US AI Safety Institutes will collaborate on research, safety evaluations, and guidance for AI safety, fostering shared capabilities through information exchange and expert personnel interactions.
The agreement, inked on Monday, April 1, 2024, builds upon commitments made at the AI Safety Summit the previous November. Technology Secretary Michelle Donelan and US Commerce Secretary Gina Raimondo endorsed the partnership to harmonize scientific methodologies and expedite the creation of comprehensive evaluation frameworks for AI models, systems, and agents.
The UK and US AI Safety Institutes have outlined strategies to establish a unified approach to AI safety testing and leverage their combined resources to effectively address associated risks. They plan to conduct joint testing exercises on accessible models and explore personnel exchanges to enhance their collective expertise.
Effective immediately, the partnership aims to streamline collaboration between the two organizations. Acknowledging the rapid evolution of AI, both governments emphasize the urgency of adopting a unified approach to AI safety to keep pace with emerging technological risks. In addition to strengthening their bilateral cooperation on AI safety, they commit to forming similar alliances with other nations to promote global AI safety standards.
UK Secretary of State for Science, Innovation, and Technology, Michelle Donelan:
This agreement marks a significant milestone in deepening the enduring special relationship between the UK and the United States to tackle the pivotal technological challenge of our era.
We have consistently emphasized that ensuring the safe advancement of AI is a collective global responsibility. Collaboration is essential to confront the risks associated with this technology and leverage its vast potential to enhance our quality of life.
The joint efforts of our nations in advancing AI safety will reinforce the groundwork laid at Bletchley Park in November, and our combined expertise will undoubtedly pave the way for nations to harness the benefits of AI securely and responsibly.
United States Secretary of Commerce, Gina Raimondo:
AI stands as the defining technology of our time. This partnership will accelerate the work of both our Institutes across various risk domains, encompassing national security and broader societal impacts. By collaborating, we are not evading these concerns but confronting them head-on. Through our joint efforts, our Institutes will enhance their understanding of AI systems, conduct more rigorous evaluations, and provide more robust guidance.
By working together, we are reinforcing the enduring special relationship between the U.S. and UK, ensuring the safety of AI both now and in the future.
The UK and the United States underscore the shared global responsibility of ensuring the safe development of AI. In recognition of the significance of ongoing international cooperation, the announcement also entails the sharing of critical information regarding the capabilities and risks associated with AI models and systems, along with fundamental technical research on AI safety and security. This collaborative effort will establish a common foundation for AI safety testing, enabling researchers worldwide to converge on a unified scientific framework.
