Report from CNBC
In Brief – UK Prime Minister Kein Starmer continued to press forward on initiatives he argues will make the UK “the best state partner” in the world for “anyone working at the AI frontier”, releasing an AI Opportunities Action Plan focused on building up computing power in Great Britain. Developed under British tech investor Matt Clifford, the plan aims to increase public sector “compute” capacity by twentyfold by 2030, including opening access to the AI Research Resource, setting up several AI “growth zones” with relaxed development regulations to speed building data centers, and forming an “AI Energy Council” with industry leaders from both energy and AI. In a Financial Times opinion piece, Starmer argues that AI technology is already helping improve lives and productivity in the UK, that AI advances are at the center of his government’s plans for the country, and that they will take a “distinctively British approach” to AI regulation that “will test AI long before we regulate” and will offer investors “stability, pragmatism and the good sense they would expect from democratic British values.”
Context – The biggest global question in AI public policy is whether governments are moving toward direct regulation or “soft law” governance such as “best practices” and “safe AI” recommendations. The EU’s AI Act is the standard for concrete regulation. And other EU regulators are stepping up action on AI. Japan, on the other hand, has been a leader of the soft law model, including through the G-7’s “Hiroshima AI Process” with principles and a code of conduct. The last US Congress did not legislate, and President Biden issued a major AI Executive Order that mostly set AI policies for the federal government but also aimed to push companies on safety. Donald Trump campaigned to repeal Biden’s order for over-regulation and the new administration is expected to push harder on AI development and regulatory restraint. In that context, a middle ground between the EU and US, such as that being described by PM Starmer, may be technically possible, but it’s not clear that it is something AI investors and developers are looking for.