UK MHRA publishes public views to guide AI health regulation

The MHRA is doing the hard, unglamorous work that will determine whether powerful AI can actually be used in healthcare at scale: listening. By running a broad consultation and publishing detailed evidence on how patients, clinicians, and vendors see AI, the agency is laying the groundwork for rules that are both politically defensible and operationally realistic. It’s also notable that this process is tied to an independent AI Commission rather than left solely to internal regulators.

For the race to AGI, healthcare is one of the few domains where frontier‑level models could both generate huge social value and catastrophic harm. The UK’s approach suggests future high‑impact deployments will come with ongoing monitoring obligations, post‑market surveillance of model performance, and strong expectations of transparency and human oversight. That won’t stop labs from building stronger models, but it will shape where and how those models can be used. Companies that can package their systems with robust clinical validation, audit trails, and adaptive risk controls will be far better positioned than those offering generic APIs. Over time, the standards hammered out here are likely to influence AI health regulation well beyond the UK.