ARTIFICIAL INTELLIGENCE CAN BE A COACH FOR LIFE AND IMPROVED HEALTH
British-based health app Babylon proves its diagnostic credentials but seeks AI solutions to keep users healthy before symptoms begin to show
It is just shy of 5pm in a London office block and the entrepreneur Ali Parsa taps on his phone. Seconds later he has a doctor’s consultation booked for 7.20pm.
“That’s the longest I’ve ever had to wait,” he says. “I’ve never seen an appointment wait this long and it’s still less than three hours.”
Just four short years after its launch, the promise of his Babylon health service as a diagnostic platform is being delivered. As a self-proclaimed industry disruptor, the former Goldman Sachs executive has set his sights on a far more ambitious goal, to ensure artificial intelligence can act as a health coach to protect users throughout their lives.
Babylon is an online doctor service that operates in Britain and Rwanda, while its AI platforms are widely available in corporate partnerships around the world.
An agreement with the Dubai Health Authority to develop the service in the UAE was signed this year. To Mr Parsa, the accord will show the promise of AI goes far beyond chatbots and video consultations.
“In the UAE there is a significant number of doctors, so why duplicate that – why go creating more doctors,” he says. “Why not add to the immediacy and the accuracy of the service. Why not put the artificial intelligence in the pocket of every UAE citizen so that they actually can have the benefits, AI, but if they need to talk to an actual doctor they can go and do that as they do now,” the former refugee says.
That is a step beyond what Babylon has offered so far. “First of all we trained the system in artificial intelligence to become the best doctor there can be,” he says.
There are obvious cost savings for patient, provider and the state. Out of 3,000 initial interactions, Mr Parsa estimates that about 1,000 led to video consultations and just 100 need face-to-face meetings with medical staff. Up to two-thirds of healthcare costs can be eaten up in salaries and infrastructure.
The trim 55-year-old, who walked across the mountains of Afghanistan as a teenager to escape his Iranian homeland, sets out the bigger challenge of catching a medical problem much earlier in the cycle. By the time symptoms present, he says, a $10 (Dh37) problem often needs a $1,000 solution.
This is where the greatest potential of artificial intelligence sits. “It learns incredibly fast. Now it needs to become the best psychologist it can be,” he says. “It needs to treat more specialties, not just do general medicine. More healthcare maintenance, how to keep you at the peak of your health, how to monitor you, how to coach you and how to predict disease before symptoms appear.
“How do we take what we know today about behavioural science and use that to get the machine to become more effective? Keeping you in compliance with your physiotherapy, with the way you should eat, with the way you should be active, with the way you should do cognitive training. We could be amazingly helpful.”
Diagnosis is a matter of probability analysis and machines are very good at it, according to Mr Parsa.
In tests, while on average doctors made the correct diagnosis 72 per cent of the time, the figure for the Babylon system came out at 81 per cent on average and 98 per cent on problems it had seen before.
With more than two million paid-up users worldwide, including 30,000 regular NHS patients in the UK, Babylon is not just an app. It operates as a call centre service in Rwanda where there is much lower smartphone penetration. A new service on Amazon’s Alexa will provide an alternative platform for the service.
In Britain’s state-funded healthcare system, Babylon faces allegations of cherrypicking the young, tech-savvy (and healthy) away from existing clinics that lose revenue while still having to treat those more prone to get sick. But to Mr Parsa it is patronising to say the elderly do not use the technology, which is more convenient than getting out of the house and down to the doctor.
The overall global approach, however, varies with a heavy reliance on partnership with local players who know the gaps in service best.
“We can’t have the arrogance to say we know best,” he tells
The National. “In each market, you can’t have the arrogance of saying, I have only one solution and that solution has to work everywhere. In the United States, we do a very different thing. In the US we are working with the health systems and the providers and employees.”
Chinese messaging platforms offer an entirely new dimension to Babylon, with a partner that already analyses every aspect of their users’ digital footprints.
“They know everything, they know every step people take,” he says. “What is the point of us trying to replicate that when they know Chinese consumers so well. Equally, it’s very hard for them to understand health care in the way we do.”
While the logical solution to problems is often simple to devise, it can be less easy to provide working solutions. “How to minimise the time of the doctors and the nurses and how to maximise the accuracy of the technology is the science behind our products,” he says.
The race to bring AI into the heart of health care will, he predicts, resemble the rapid concentration of the personal computer industry four decades ago.
“When personal computers came along, it was a hobbyist industry ... there very few global suppliers,” he said. “When we first started there were so many contenders and already today the number is very low. People look under the bonnet, they see who has depth and who doesn’t.
“As long as somebody makes healthcare access more affordable and puts it in the hands of every human being on the earth, I’m happy.”
While doctors on average made a correct diagnosis 72 per cent of the time, the Babylon system came out at 81 per cent