The United States and Britain have announced a new partnership focused on the safety of artificial intelligence (AI) technology, amid increasing concerns about the upcoming next-generation versions.
Commerce Secretary Gina Raimondo and British Technology Secretary Michelle Donelan signed a memorandum of understanding in Washington to collaborate on advanced AI model testing. This agreement follows commitments made at an AI Safety Summit held at Bletchley Park in November.
Raimondo emphasized the significance of AI as a defining technology of our era, stating that the partnership will accelerate efforts to address national security risks and broader societal concerns related to AI.
Both countries have established government-led AI safety institutes. Britain announced the establishment of its institute in October, while the United States launched its own in November. The US institute is focused on evaluating risks associated with frontier AI models and is collaborating with 200 companies and entities.
Under this partnership, Britain and the United States aim to conduct joint testing exercises on publicly accessible AI models. They are also considering personnel exchanges between their institutes and plan to develop similar partnerships with other countries to promote AI safety.
Donelan highlighted the importance of addressing AI risks, emphasizing that while AI has great potential to address global challenges, it is essential to manage associated risks effectively.
Generative AI, which can create text, photos, and videos in response to prompts, has generated both excitement and concerns regarding its potential impact on jobs, elections, and human autonomy.
Raimondo and Donelan stressed the urgency of joint action to address AI risks, particularly with the imminent release of more capable AI models. Raimondo also mentioned plans to raise AI issues at a meeting of the US-EU Trade and Technology Council.
Both countries plan to share key information on AI capabilities and risks, as well as collaborate on technical research related to AI safety and security.
In October, President Biden signed an executive order aimed at reducing AI risks, while in January, the Commerce Department proposed regulations requiring US cloud companies to assess whether foreign entities are accessing US data centers for AI model training.
Britain announced plans in February to invest over GBP 100 million to establish nine new research hubs and train regulators on AI technology.
Raimondo expressed particular concern about the potential misuse of AI in bioterrorism or nuclear war simulations, emphasizing the need for stringent measures to prevent catastrophic consequences.