Advancing AI Regulation in Healthcare: Insights from AI Airlock Phase 2

The Medicines and Healthcare products Regulatory Agency (MHRA) in the UK has released the report for Phase 2 of its AI Airlock regulatory sandbox, a collaborative initiative designed to explore and refine the regulation of AI medical devices. This phase, which ran from April 2025 to March 2026, involved working closely with seven diverse AI technologies, ranging from advanced cancer diagnostics to AI-powered clinical note-taking systems. The program aimed to address critical regulatory challenges inherent in the rapid evolution of AI in healthcare. One of the primary focuses of Phase 2 was managing the intended purpose of AI products throughout their lifecycle and validating generative AI medical devices. It also delved into ensuring safe lifecycle management through predetermined change control plans (PCCPs) and robust post-market surveillance (PMS). Furthermore, the program sought to strengthen performance evaluation approaches for AI-powered in vitro diagnostics (IVDs). The findings from this phase strongly advocate for a more dynamic and lifecycle-oriented approach to regulating AI medical devices. The MHRA acknowledges that pre-market validation alone is often insufficient for AI medical devices, as their real-world performance can be difficult to replicate in controlled environments. This highlights the critical need for robust post-market monitoring. Consequently, pre-market evidence requirements must be designed with deployment conditions in mind, and post-market oversight activities should be planned early, complementing the technological and lifecycle needs of the device. Key recommendations emerging from Phase 2 include the development of clear guidance on PCCPs specifically for AI medical devices. Additionally, there is a call to update existing software qualification and intended purpose guidance to incorporate AI-related considerations. The MHRA also recommends establishing specific performance evaluation guidance tailored for AI-powered IVDs. These insights are not merely theoretical; they are actively contributing to the evidence base for the National Commission into the Regulation of AI in Healthcare and will continue to inform both UK and international guidance for AI medical devices, ensuring that regulatory frameworks keep pace with technological advancements.