Written evidence submitted by Nicole Whippey (DCG0003)

Having worked in medical device software development at Renishaw Neuro Solutions for almost a decade. More recently I have supported IngeniumAI, a medical device AI start up, that was a spin off from Bath University. Their AI research received funding while in academia. However the steps to move AI from research to a regulated medical device, a requirement for clinical deployment, seems to be overlooked often. It requires a huge amount of quality/regulatory knowledge, as well as good software life cycle development practices which is not the norm for researchers in universities. To add to this, an AI medical devices requires clinical user needs, requiring the specific clinical workflow to be defined. AI solutions are so context specific, and without this need up from, then a whole host of AI clinical solutions could have time and money spent researching them, when they don’t work at all for either those using them, or within the specific context of use. I urge you to not only have an Incubator for AI, but also a translational hub for AI in regulated industries. For medical devices this would require quality/regulatory experts, experts in clinical evaluation, software development life cycle managers. In an ideal world, a service that works from research inception and helps with user needs and clinical workflows, to being able to move research through rapidly to meet regulatory requirements. What is the point of investing loads of money in AI research if it never gets translated into clinical use? The researchers are not the ones with those skills, so it makes sense for DSIT to be responsible for making real world difference with the research it is investing in.

12 February 2025