Report from the Financial Times
In Brief – UK AI Minister Kanishka Narayan has urged frontier AI companies not to release models they cannot adequately control, while calling for Britain to develop sovereign infrastructure that gives the state greater control over emerging technologies. Narayan is developing plans that could legislate mandatory safety reporting, monitoring of advanced AI systems, and potentially require companies to provide pre-release access to models. He also confirmed that the UK’s AI Security Institute (Aisi) would regain access to US-developed models for testing after Anthropic previously declined a request following pressure from the Trump administration. Narayan said discussions had persuaded Washington not to block future sharing. Comparing AI’s emergence to the Industrial Revolution, Narayan argued that legislation and enforcement institutions were necessary to prevent harmful outcomes, and while acknowledging public anxiety about massive economic and security risks, he identified cybersecurity and biosecurity as more immediate concerns than existential AI risks.
Context – Over the past five years, while the UK has enacted and started implementing a digital regulatory regime that parallels the EU’s, government leaders often pitched that the UK was more predictable and digital business friendly. Although the government of PM Rishi Sunak pushed to the forefront of AI “Safety” discussions, where the EU enacted its AI Act, the UK did not. The Trump Administration’s hard push against AI regulation tamped down regulatory urges nearly everywhere, and even the EU may dial back some, but the pendulum continues to swing. The UK Parliament’s Joint Committee on Human Rights has proposed an AI Bill and national regulator to carryout pre-release testing, block or withdraw high-risk systems, investigate harms, enforce safety standards, and require AI users to disclose its use in significant decisions. At the same time, despite widespread AI angst, the IMF’s latest global growth forecast upgraded its UK economic projection, saying it was emerging as a winner of the AI and tech investment and services boom.