Generative AI and Large Language Models (LLMs) have the potential to play a “transformational” role in the evolution of healthcare, helping patients to take responsibility for their own health care and reducing demands on a stretched healthcare workforce, but a strong regulatory framework is essential, Umang Patel, chief clinical information officer at Microsoft, said. 

Patel was speaking to Digital Health News ahead of Digital Health AI and Data, a new event running 30-31 October 2023 at London’s Business Design Centre.

He will share the AI and Analytics stage with Professor James Teo, professor of neurology at King’s College Hospital on Day One of the event, presenting a Masterclass on the future of generative AI in health and care. 

Patel, a trained pediatrician, has already experienced the changeable fortunes of the AI start-up world, as “employee number 10” at digital health start-up Babylon Health, a company that was one of the first to offer virtual appointments and chatbots, but which recently put its UK business up for sale and filed for bankruptcy for two US businesses.

Having left the company just after its IPO, Patel said he took away the lesson that although new business models for healthcare are worthwhile, it is important to do them at scale. 

“We bit off more than we could chew at Babylon,” he reflects. “The things that defined our success were tempered by the fact that the US market is entirely different and consumed too much money and resources. The health system is set up so much for service provision as opposed to prevention that it makes it very difficult to make the economics work.” 

“I’m confident that some of the advances we are seeing in large language models will provide resources for people to help care for themselves,” he added. “The only renewable energy in healthcare is patients. We need to better allocate resources and shift into a model of identifying people before they are sick.” 

No single breakthrough 

While he is confident about the potential of new healthcare technologies, Patel says he doesn’t see one development that is likely to have an outsized impact, such as the invention of magnetic resonance imaging (MRIs). 

“The impact will come from the fact that we can bring consumer grade technology into healthcare; general purpose AI will mean that we can start thinking about things in an entirely different way,” he said, adding that AI is likely to take an  assistant role in the healthcare sphere, similar to a co-pilot on a plane. 

“Without help from a co-pilot, you couldn’t fly long distances, but you would never get into a plane that didn’t have a pilot in it… I think that’s where we are with healthcare”, he said.

In terms of extending existing healthcare capabilities, generative AI is likely to be most valuable in disease areas with the greatest need, such as mental health, Patel said, helping to tackle loneliness and provide people a safe space to talk without feeling that they are judged. 

Secure regulatory regime is key 

Regulation and support will be essential to the expansion of LLMs in healthcare, Patel said, adding that he is optimistic about the prospects for global cooperation on regulating AI. He used the analogy of electricity which operates safely in most parts of the world, despite differences in voltage.  

Healthcare professionals need to be involved in the regulation process and be willing to engage in the development of AI so they can understand it enough to know how to build the networks to run it, Patel said. 

He concluded: “We need to collectively come together to understand what we want for the technology and start setting it up; it will be foundational.” 

AI and Data is from the organisers of the market-leading Digital Health Rewired and Digital Health Summer Schools events.  

All sessions are CPD accredited, and the full programme agenda can be found here. AI and Data is free for the NHS, public sector, start-ups, charities, education and research. Commercial tickets start from £275+VAT. Register here.