The UK’sMedicines and Healthcare products Regulatory Agency (MHRA) just dropped a report that should concern everyone betting on AI to transform the NHS. Their National Commission on the Regulation of AI in Healthcare ran a call for evidence from December 2025 through February 2026, and the responses were… let’s say, not subtle.

Seventy-one percent of respondents disagreed or strongly disagreed that existing regulation is sufficient for safety and performance standards. That’s not a marginal critique. That’s a near-universal rejection of the current framework.

The survey drew 761 unique responses from NHS bodies, industry, academia, and patient groups—real stakeholders who actually build, deploy, and live with AI medical devices in the NHS today. Their concerns cluster around four areas: safety, performance standards, data governance, and transparency. But the most striking finding might be this: 50% said the existing framework needed substantial revision, while 21% called for a complete overhaul. That’s 71% of respondents essentially saying “this isn’t working.”

The current system regulates AI in healthcare under the UK’s medical device regulation (MDR), classifying the technology as software-as-a-medical-device (SaMD) or AI-as-a-medical-device (AIaMD). But as AI systems have grown more powerful and more embedded in clinical workflows, the regulatory framework hasn’t kept pace. Respondents highlighted that post-market surveillance for AI health technologies is particularly weak—once an AI system is deployed, who’s monitoring whether it’s still performing as intended? Who’s ensuring healthcare professionals have the training to use it? These are basic questions that the current framework doesn’t answer adequately.

The National Commission’s final recommendations are due in the coming months, and they’ll inform a new regulatory framework aimed at making the NHS “the most AI-enabled healthcare system in the world.” That’s an ambitious goal embedded in the NHS 10-year plan. But here’s the tension: you can’t become the world leader in AI healthcare with regulatory frameworks that 71% of your stakeholders think are inadequate.

The lesson here isn’t that the UK should abandon safety oversight. It’s that ambitious AI adoption goals require ambitious regulatory reform. You can’t ask the NHS to embrace AI at scale while running a regulatory system that the people building and using the technology have lost confidence in.

The MHRA has an opportunity to lead—not just in the UK, but globally, where there’s “absence of clear consensus” on how to regulate AI in healthcare. What they do next will set the tone for whether the UK captures the productivity benefits of AI in healthcare or watches other countries move faster with clearer rules.