The jargon-free guide to AI and health
Can artificial intelligence really keep patients out of hospitals? Keith Grimes, Clinical Artificial Intelligence & Innovation Director at Babylon Health explains how AI works.
There’s so much hype and confusion about the term, but to me artificial intelligence is just the use of a computer to do something that might otherwise require a human brain. But instead of building a replica of a human brain, we’ve developed a whole range of individual tools.
For example, computers have got very good at understanding images, so they can do handwriting recognition or facial recognition better than humans.
Mostly, they’re very good at recognising patterns. If you have huge amounts of data and huge amounts of processing power, they can recognise patterns that might otherwise not be visible to the human eye, and certainly not the untrained eye.
Of course, you often need an enormous amount of data to do this. It needs to be good quality and labelled or categorised by humans to accurately train the machines.
Sometimes people say to me “I don’t like AI, I don’t trust it, I’d never use it.” But AI tools are everywhere today. If you’ve ever used the GPS on your phone or in the car, you’ve used an artificial intelligence system trained on a vast amount of data to find the quickest route between two points. Once AI becomes sublimated into our world, we don’t think about it or even see it.
When you speak to a GP, particularly in the UK, their clinical systems will normally include a prescribing module. This is a relatively simple system, powered by huge amounts of data, that helps me every time I prescribe medicine for a patient. In the past, I’d have relied on my training and the British National Formulary, a 1,700-page book of data about medicines.
So where is this going?
This isn’t science fiction, it’s just looking at what people are doing now and extrapolating out 3-5 years. Today, homes are filling up with more and more sensors. Not specialist medical devices, but things like Amazon Alexa and Google Home. They are becoming more powerful, and more able to pick up measurements that are important for healthcare, things like activity patterns.
Then there’s wearable devices like Apple Watch that can sense step count, pulse, oxygen saturation. Over the next five years they will probably be able to measure blood glucose levels. If you think people like tracking their steps, wait until they can track blood sugar!
A lot of healthcare is all about confidence and trust. There’s been countless times when my arrival on the scene as a doctor has defused a patient’s anxiety just because they trusted me. All I’ve done is spoken to people and nothing else. I believe that machines can help build that trust.
Keith Grimes - Clinical Artificial Intelligence & Innovation Director at Babylon Health
With all this data, and with really strong human services to back it up, it will be possible to help patients and families to develop the trust and confidence to stay at home and avoid a trip to the doctors or a routine hospital appointment.
A lot of healthcare is all about confidence and trust. There’s been countless times when my arrival on the scene as a doctor has defused a patient’s anxiety just because they trusted me. All I’ve done is spoken to people and nothing else. I believe that machines can help build that trust.
There’s still a lot to do. For every group, every condition, we need to collect the data, develop the processes, and gradually build up that confidence so that in the right circumstances, more people can stay at home, which is better for everyone.
