The UK is going its own way on AI regulation in health, but problems lie ahead

Abstract

It is a frenzied time for UK health policy. NHS England is set to be abolished, amid an emerging 10 year plan, and every UK country faces difficult financial situations. This is set against two historic global shifts that bring deeper, slower change. Firstly, a “reset” in UK-EU relations brings the promise of deeper cooperation on security, agriculture, trade and travel. Secondly, the ever-faster emergence of artificial intelligence (AI) and governments’ scramble to respond. But as the UK draws closer to the EU in crucial areas, it is going the other way on AI.1 While the EU has passed an AI Act for comprehensive regulation, on AI the Labour government is so far sticking to the same approach as its Brexit-oriented predecessor—claiming the UK’s “pro-innovation approach” is “a source of strength that should be preserved.”23 These differences and the pace of technological change mean difficulties ahead, with no easy answers. #### The divergence dilemma The technical standards for AI as a medical device are similar in the UK and the EU. However, the EU requires a Fundamental Rights Impact Assessment to be conducted, unlike the UK, and those technical standards may diverge further in future. The EU applies a single framework to every use of AI, while the UK asks each individual …