MHRA launches : The AI- Airlock ! (And will it break the evidence barrier)

Whenever I think of the word airlock, I have a scene from 2001 : Space Odyssey play in my mind where HAL ejects Dave - saying I'm sorry, Dave.... how chilling. (Let's all hope that isn't where this is all heading :D)

So today we have news of a different kind of AI airlock:

MHRA to launch the AI-Airlock, a new regulatory sandbox for AI developers - GOV.UK (www.gov.uk)

Anyone who's been trying to develop AI in healthcare has come up against the evidence barrier. "But sir, my product can't really be run as a RCT, it doesn't work that way... well then you wasted your time. Don't bother me. You know that's how we do things in the good old NHS. If only the Queen were still alive. She'd give you what for. I'm off to play golf...."

(Fictional conversation between Healthtech and prof.)

So in answer to this conundrum :

The Medicines and Healthcare products Regulatory Agency (MHRA) has announced the launch of a new ‘regulatory sandbox’ called the AI-Airlock.

This sandbox will provide a regulator-monitored virtual area for developers to generate robust evidence for their advanced technologies.

As we all know AI has massive potential, but these technologies can sometimes be challenging to test using traditional trial techniques. The AI-Airlock project aims to address this by offering a collaborative approach to identifying and managing evidence requirements.

Brilliant!! But how?!

This initiative will allow patients to benefit from faster access to developing technologies such as improved diagnostics or precision medicine. It will also enable advanced AI technology to be used in NHS settings, with strict safety controls, ahead of navigating regulatory approval.

The AI-Airlock is designed to be a collaborative space, bringing together expertise from innovators, regulatory organizations including Approved Bodies, Government, the NHS, and academia. The service is set to launch in April 20242.

So what does this mean?

On the box it says that the aim is to bring cutting-edge products to UK patients faster, without compromising on its robust standards of safety and performance.

The process follows the ‘regulatory sandbox’ model which is a world-leading mechanism to assist in safe development and deployment of software and AI medical devices.

This learning can then be shared, helping to provide an evidence base that promotes a wider understanding of the challenges and potential solutions that are available.

This initiative will support innovators to work within the current regulatory system, identify where their products need to build more evidence needed for a safety and efficacy assessment and help resolve these issues. It follows a robust process, so manufacturers of software and AI medical devices understand and deliver what is required to ensure the real-world viability of these devices.

It sounds brilliant, tell me more about this sandbox business?!

Right then, a regulatory sandbox you say?

A regulatory sandbox is a framework set up by a regulator that allows businesses to test new innovations in a controlled environment under the regulator’s supervision.

The purpose is to facilitate innovation in a safe and responsible manner and to see if the products can enter the market.

A regulatory sandbox usually involves relaxed regulatory conditions, a limited number of customers, and a limited period of time.

This approach allows innovators to trial new products, services, and business models without some of the usual rules applying.

The sandbox is offered by regulators like the Financial Conduct Authority (FCA) and provides firms with access to regulatory expertise and a set of tools to facilitate testing.

The sandbox enables innovators to understand what rules might mean for them and where their products need to build more evidence needed for a safety and efficacy assessment. It helps resolve these issues and supports firms in conducting live tests with the regulator’s oversight.

In Summary then:

In summary, a regulatory sandbox is a safe space where innovators can test their novel ideas under the supervision of regulators, ensuring that they comply with necessary regulations while also allowing for creativity and innovation.

It sounds brilliant, and I'm really looking forward to seeing what happens in April 2024, as that will (hopefully) be 4 months before my PHD starts when I will be absolutely in need of some proper frameworks to develop an AI product in the NHS.

Have a brilliant day, Dave.