The role of artificial intelligence (AI) in healthcare, so far anyway, has typically been in automating paper-based processes. This may soon change, as AI becomes central to diagnostics, treatment and even surgery.
Boris Cergol, head of AI at Comtrade Digital Services, said that AI in healthcare is following the same trajectory as in other fields: in short, it is working its way in from the fringes to the centre of activity.
“What I think is happening in healthcare and, in fact, is happening in application of AI systems, is it is being moved from being applied to information databases to being applied to more of a real world setting,” he said.
“What I mean by that is [that] sensors are present in the environment, which profile information in real time, more and more robotics applications and, of course robots interacting with the [actual] physical environment,” he said.
Indeed, much of the discussion about ‘robotics’ in recent years has been about software: robotic process automation.
What Cerjol is proposing, however, is actual robots interacting with patients.
Comtrade works across the domain of AI including in health and the pharmaceutical sectors and Cerjol said that AI moving into healthcare is all part of a wider move away from simply automatic paper-based processes.
AI systems are now gaining traction in developing new innovations, for example, finding substances that may have medical uses or working through research to suggest directions.
“These are things [that] I find very encouraging,” he said.
“We are moving from analysing structured data [such as admission records] to less structured data. That could be ultrasound or X-ray images, some comments written about patients [or] a recording from a doctor. This kind of inflation thaw was really difficult to analyse in the past.”
Talking about healthcare right now is difficult to do without raising the spectre of the SARS-CoV-2 novel coronavirus. This poses the question, though: AI is all very interesting, but what has it done for us lately?
Cergol said that while it is true that AI has not made an enormous contribution, that will change in the coming years.
“I do think that there’s an overall impression that AI has failed in relation to Covid-19 among the wider population, but in order to understand what is happening you have to look at the detail. At the core of the crisis was [the fact] that we didn’t have enough capacity in our healthcare systems.”
This means data applications of AI had little contribution to make as the problem was never one of data.
“The Covid crisis did not introduce a new information system problem,” he said.
Despite this, Cergol said that progress in real world applications of AI is so fast that any future pandemics may be very different.
“If we went through this kind of crisis five to ten years [hence] we would have been able to respond very differently," he said.
Cerjol said the goal is to use AI not to replace human actions and intelligence, but to support and augment them – including in the field.
“We need to see if we can help with care for patients by using robotics or analysis of vital signs. This is to make the work of the people in the front line easier,” he said.
In fact, he said, this is part of a three-stage process in the workplace.
“We’re well into the first, which is decision support though datasets and electronic health records and so on. The next will be replacing some of the more repetitive work.”
Following this will be the ‘interfacing’ of AI with the physical world.
“The third stage is the introduction of robotic processes. We see this deployed in kitchens, and it will be interesting to see similar systems deployed in healthcare.”
One area that will take time will be mental health.
“It’s an area that needs to be approached carefully,” he said.
Cerjol said that he does not expect to see AI replacing the conversations people have with therapists.
“However,” he said, “it will be possible to alleviate some of the loneliness that people feel by having conversations. Conversational systems will not be simply reactive, but they can help us by reminding us to do things, provide high-quality information and so on.”
Cerjol said that this will also be driven as much by demographics as by the technology itself: an ageing population requires more healthcare.
“Almost everywhere in the world, healthcare is understaffed and the need for it will be rising. People need not be afraid that they will be replaced by robots,” he said.