Despite a recent joint initiative to assess risks from artificial intelligence, the United States and the United Kingdom will maintain their pro-innovation approaches to AI that will keep regulations less robust than those in the European Union, although London's AI laws may grow stricter after the country's upcoming general election. The United States and the United Kingdom signed a first-of-its-kind agreement on AI safety on April 1, becoming the first countries to formally initiate bilateral cooperation on AI testing and risk assessments for emerging AI models. The agreement lays out how the two governments will share technical knowledge, information and talent on AI safety, as well as how they will evaluate private AI models via the United Kingdom's new AI Safety Institute, or AISI, and its U.S. counterpart of the same name. The bilateral agreement and the formation of these two institutes follow the November 2023 AI Safety Summit, which the United Kingdom hosted. During the summit, more than 25 countries — including the United States and China — and the European Union signed a voluntary commitment to AI oversight and safety advancement called the Bletchley Declaration. 

• The Bletchley Declaration details a two-part agenda for signatory states that focuses on identifying AI-related safety risks and developing cross-country risk-based policies to ensure the safe development of AI tools. Private sector representatives at the event included OpenAI's CEO Sam Altman and X's CEO Elon Musk. 

The announcement and other recent initiatives reflect the U.S. and British governments' shared approach of minimally regulating AI. Neither the United Kingdom nor the U.S. federal government currently has AI-specific legislation in place that regulates its use in the commercial and private sectors. Instead, they rely on a patchwork of existing regulations such as consumer and data protection requirements, as well as product safety and transparency conditions. The lack of a more extensive framework in the two Western countries is due in part to the current British and U.S. governments' efforts to ramp up the competitiveness of their AI sectors via less stringent oversight protocols. The British government, under the current leadership of Prime Minister Rishi Sunak and the Conservative Party, has sought to boost AI investment in the country by reducing regulatory requirements. In February, for instance, London published a consultation response to a white paper detailing updated information on plans to regulate AI in the country via a context-based and sector-specific approach that relies on preexisting rules to oversee AI applications. Meanwhile, in the United States, President Joe Biden's administration recently passed updated AI guidelines for federal agencies and public sector entities, but bipartisan leaders have broadly agreed that AI regulation should be taken slowly, stymying the development of robust legislation on AI regulations and use cases in the private sector. Rather, the Biden administration has relied on unilateral efforts like executive orders and voluntary, nonbinding private sector commitments on AI safety.

• Major AI developers including OpenAI, Google DeepMind, Microsoft and Meta have signed voluntary commitments to open up their latest generative AI models for review by the United Kingdom's AISI once it has devised a review process for advanced models. However, AISI's reviews do not legally require these companies to change their product releases based on the outcome of the tests.
• In July 2023, seven AI companies — Amazon, Anthropic, Google, Inflection AI, Microsoft, Meta and OpenAI — publicly announced voluntary commitments in partnership with the White House to develop safety and transparency tools to minimize security risks in their AI products. 

While Washington and London's coordination on AI safety may bolster some AI oversight, their limited regulatory approaches to AI regulation mean these efforts will be much less robust than frameworks in jurisdictions like the European Union. The formation of the AISI and future collaboration between the two countries and private sector entities will likely enhance AI oversight efforts in the future. Nevertheless, the British and U.S. governments' attention on international AI competition and adversarial states like China suggests that they want to maintain a more relaxed attitude toward AI regulations, shying away from robust, intensive laws for commercial entities and instead relying on preexisting laws and limited enforcement on AI use cases. The two governments are also likely to further invest in private sector relations and bolster collaboration with top AI company representatives, which may contribute to further voluntary and nonbinding commitments by these companies but are unlikely to result in the creation of more enforceable rules. While these efforts will nurture the AI sector's growth and advancement, the British and U.S. governments' pro-innovation approaches to AI regulation will also be much weaker than more robust legislative frameworks in jurisdictions like the European Union, which adopted its Artificial Intelligence Act on March 13. Though the European Union's AI framework is expected to pose challenges for companies, it is also slated to offer consumers much more protection from current and future AI tools and use cases. As a result, while AI startups and other tech companies are likely to benefit from the ambiguity in British and U.S. regulations, consumers will be less protected and may be more likely to have their data co-opted or misused for commercial purposes.

• In its consultation response, the British government claimed it would only legislate new AI risks if it determined that existing mitigation measures were no longer adequate and if the government was not sufficiently confident that all relevant parties would effectively implement all voluntary measures.
• The EU AI Act, which fully goes into effect in mid-2026, applies to all AI systems and requires a range of new transparency and reporting requirements, including risk assessments and public disclosures around AI training and implementation procedures. 

The United Kingdom's approach to AI is subject to change depending on the outcome of the next general election, which the center-left Labour Party is currently projected to win. The next British general election is widely expected to take place before the end of the year, and current projections suggest that Sunak's Conservative government will lose power, which would usher in a new Labour Party-led government and a likely change to AI oversight initiatives. Though the Labour Party has not yet released an AI roadmap, Labour lawmaker Peter Kyle stated on March 12 that the party was currently developing an AI strategy document and would seek to establish a government body to create clear and consistent regulations for the tech and AI sectors. Although Labour Party representatives have emphasized the importance of leveraging the capabilities of AI tools in the British economy and marketplace, the party has also hinted at putting controls on an array of AI tools including generative AI products that can create misleading or fictitious content. The Labour Party's official approach to AI is still in development and will likely also attempt to bolster innovation in the country in line with the Tories' objectives but will ultimately likely prioritize regulatory oversight regarding these nascent technologies. As a result, the British government's long-term outlook on AI may slide toward a more EU-like approach, though the Conservative and Labour parties' combined objective of advancing the United Kingdom's AI sector means the country is unlikely to go as far as the European Union, even under a Labour government.

• According to the British election prediction site Electoral Calculus, the Labour Party has about a 96% chance of winning a majority based on current opinion polls.
• Following Sunak's November 2023 claim that he would not "rush to regulate" AI, the Labour Party put out a statement claiming it would introduce binding requirements for companies developing AI systems.
• In February, Labor lawmaker Kyle also claimed that the Labour Party would require AI firms to share the results of their testing and risk assessments, replacing Sunak's current voluntary testing agreement with tech companies and creating a statutory code for all companies developing AI tools. 
• On March 17, the Labour Party-aligned think tank Labour Together published a policy paper proposing a legal ban on certain generative AI content like deepfake technology and disinformation generators.