🇬🇧 UK Selects 7 AI Medical Tools for MHRA Airlock Clinical Trials: Is Smart Licensing Finally Here?

🟨 Introduction: From Experimentation to Licensing—Have the Rules Changed?

In the world of digital medicine, it’s no longer enough for a tool to be smart or accurate—it must be safe, legally compliant, and clinically usable. In a groundbreaking move, the UK’s Medicines and Healthcare products Regulatory Agency (MHRA) announced the selection of seven AI-powered medical tools for phase two of its experimental program, Airlock. This initiative aims to test these tools in a controlled clinical environment before granting full regulatory approval.

These aren’t just data analysis tools—they’re language models, diagnostic algorithms, and imaging technologies powered by AI. Some can detect cancer in minutes, while others interpret blood results instantly.

So, are we witnessing the dawn of smart licensing? And will AI tools become part of the medical decision-making process in the UK?

🧠 What Is the MHRA Airlock Program?

Launched in 2023, Airlock is a regulatory initiative by MHRA designed to test AI medical tools in a closed clinical setting before granting full approval.

The name “Airlock” is inspired by space station protocols, where equipment is tested in isolation before entering the main system. In medical terms, it means the tool is used under supervision, without yet being fully licensed.

The program’s goal is to accelerate the adoption of medical innovation without compromising safety, ethics, or regulatory standards.

🩺 The 7 Selected Tools: From Imaging to Voice Recognition

In October 2025, MHRA announced the selection of seven AI medical tools from over 80 applicants. These tools span a wide range of clinical applications—from image analysis to voice-based diagnostics—and are now undergoing rigorous testing within the Airlock framework.

Qure ai : Specializes in analyzing chest and head X-rays. In emergency settings, it can detect brain hemorrhages within seconds, enabling faster intervention. Trials show up to 96.4% accuracy in critical cases, with deployments in several Asian countries.

Lunit INSIGHT : Focuses on breast cancer detection via mammography. Trained on millions of images, it reduces human error by 34%. It’s already in experimental use in South Korea and the U.S., and now under MHRA review.

Kheiron Medical :
A UK-based company offering a breast image analysis algorithm trained on over 3 million scans. Previously supported by the NHS, it’s a strong candidate for full licensing.

Corti AI : A voice-based tool used in emergency calls. It detects signs of cardiac arrest from speech tone and word patterns, reducing response time by an average of 12 seconds. Currently being trialed in Denmark.

DeepMind Health (Google) :
Developed a language model to analyze medical records and suggest treatments. Shows over 90% accuracy in summarizing complex cases, making it one of the most advanced tools in the program.

Skin Analytics : Uses visual analysis to detect skin cancer via smartphone or imaging device. Backed by NHS in early trials, it’s now undergoing broader evaluation.

FibriCheck : A mobile app that detects heart rhythm disorders using a phone camera and PPG technology. Partially licensed in Europe, now under full review in the UK.

📊 Program Goals: Regulation That Doesn’t Kill Innovation

Airlock isn’t just about testing tools—it’s about redefining the relationship between innovation and regulation. MHRA wants to ensure that AI tools deliver real clinical value, don’t expose patients to uncalculated risks, and comply with ethical and legal standards.

Over 80 tools were submitted for phase two, and only seven were selected based on strict criteria including accuracy, safety, and compliance. The testing period ranges from 6 to 12 months, overseen by an independent panel of doctors, engineers, and ethics experts.

🔐 Privacy & Ethics: Is Patient Data Safe?

Privacy is a core pillar of Airlock’s evaluation process. MHRA requires that:

Data is processed in a closed environment without permanent storage
No data is used for commercial or external training purposes
All tools comply with GDPR and the UK Data Protection Act

Any tool that violates these standards is immediately disqualified, regardless of its technical performance. This reflects MHRA’s commitment to ensuring AI serves patients—not exploits their data.

🧬 Can These Tools Receive Full Licensing?

Yes, but full licensing isn’t automatic. It requires passing a series of rigorous regulatory stages, starting with documented clinical results that prove the tool performs safely and accurately in real-world medical settings.

Each tool must demonstrate not just precision, but consistency across diverse cases and clinical conditions. It must also integrate smoothly with hospital systems without adding operational risks or complexity.

Once these conditions are met, the results are submitted to an independent MHRA panel comprising experts in medicine, data science, ethics, and law. This panel conducts a comprehensive review of the tool’s safety, effectiveness, and ethical compliance. Only after passing this review can full licensing be granted.

Notably, in Airlock’s first phase, only one tool received full approval after 11 months of evaluation—highlighting the program’s strict standards. This approach doesn’t aim to slow innovation, but to ensure that every tool entering the UK healthcare system is trustworthy and clinically valuable.

🌍 Can These Tools Be Used Outside the UK?

Some tools, like Lunit and FibriCheck, already hold partial licenses in Europe and Asia. MHRA is working to develop a licensing model that could be adopted globally, in collaboration with the European Medicines Agency (EMA) and the U.S. FDA.

This means success in Airlock could pave the way for international adoption—provided the tools meet local regulatory standards.

❓ Frequently Asked Questions

① Are these tools available to the public now?

No, they’re only used in closed clinical settings under MHRA supervision.

② Can doctors rely on them entirely?

No, they serve as decision-support tools. Final decisions remain with the physician.

③ Do these tools learn from patient data?

Yes, but only within a closed environment without permanent data storage.

④ Can they make mistakes?

Yes, like any analytical tool, their accuracy depends on data quality and context.

⑤ Can they be used at home?

Some, like FibriCheck, are designed for home use but require strict evaluation before full approval.

⑥ Can they be commercialized?

Not within Airlock. Commercial use is only allowed after full licensing.

⑦ Will more tools be added later?

Yes, MHRA has announced that phase three applications will open in January 2026.

🟩 Conclusion: The UK Isn’t Just Testing Tools—It’s Redefining Medical Licensing

What’s happening in Airlock isn’t just a technical trial—it’s a regulatory revolution. In a world of accelerating innovation, regulators are no longer relying on paperwork alone. They’re testing tools in real clinical environments and redefining what “approval” means.

AI won’t replace doctors—but it will become part of their decision-making. And licensing won’t be granted based on promises, but on proven results.

Is this the beginning of a global model? It seems the UK is drawing the blueprint.