When your doctor’s diagnosis comes from an algorithm, the question of who regulates that algorithm suddenly matters a lot. Governments on both sides of the Atlantic are now scrambling to answer it.
The UK’s National Commission into the Regulation of AI in Healthcare dropped 44 recommendations on September 9-10, 2026, calling for a fundamental overhaul of how AI-powered medical devices get approved and monitored. The US Food and Drug Administration released its own discussion paper on generative AI medical devices in August. And the European Union’s AI Act is rolling out high-risk classifications for medical AI systems over the next two years.
The UK wants to tear up the old approval model
The UK commission, established by the Medicines and Healthcare products Regulatory Agency (MHRA) in September 2025 and led by NHS doctors, is proposing something genuinely different. Instead of the traditional approve-it-once-and-forget-it model, the commission wants a lifecycle-based, risk-proportionate system.
One of the more interesting proposals is what the commission calls “L-plate” authorizations, essentially staged approvals where AI systems operate under supervision before earning full clearance. The framework also calls for continuous monitoring of real-world performance, public access to safety data, and beefed-up enforcement powers for the MHRA.
The commission gathered input from over 12,000 stakeholders. Doctors, patients, tech companies, and regulators all weighed in on what responsible AI deployment in healthcare should look like.
The FDA takes a cautious first step
Across the Atlantic, the FDA’s August 18, 2026 discussion paper focuses specifically on generative AI-enabled medical devices, laying out a competency-based framework that covers risk assessment, premarket evaluation, and postmarket monitoring. Public comments on the paper are due by October 19, 2026.
The FDA hasn’t actually authorized any generative AI-enabled medical device yet. It did grant breakthrough designation to one patient-facing clinical application involving generative AI back in March 2026, a designation that fast-tracks review but doesn’t equal approval.
Meanwhile, individual US states aren’t waiting for federal regulators to sort things out. Colorado enacted a healthcare AI law effective August 12, 2026, mandating human oversight for AI used in therapy and psychotherapy applications. Multiple other states passed similar legislation in 2026, all centered on the same principle: a human should remain in the loop when AI is making decisions about someone’s mental or physical health.
Europe plays the long game
The EU AI Act classifies many medical AI systems as high-risk, which triggers a suite of compliance obligations around transparency, data governance, and human oversight. Full high-risk obligations won’t be enforced until December 2027 for standalone AI applications and August 2028 for AI embedded in specific devices.
In January 2026, the European Medicines Agency (EMA) and FDA jointly published ten guiding principles for AI in medicine development. The principles emphasize ethical AI use, clear accountability structures, and safe application throughout a product’s lifecycle.