AI and the National Health Service
It’s no secret that the National Health Service is in dire straits. Underfunded, understaffed and oversubscribed, its continuous struggle to provide care to Britain’s growing population is well documented. Now, some experts are claiming that AI – and cognitive agents in particular – is the only way to save the service.
Using cognitive agents to improve care
So-called cognitive agents – Artificial Intelligence that is able to hold a human-like conversation – is considered by some to be the key to rescuing the NHS and providing the care levels that citizens expect.
By training ‘digital employees’, says IPsoft’s Global Cognitive Health Solutions director David Champeaux, the pressure on doctors and nurses would be partially relieved. The cognitive agents could tackle the routine and administrative tasks that currently take up so much of medical employees’ time, leaving them with more capacity for patient care.
The past winter saw almost 55,000 operations cancelled due to increased pressure on ambulances and A&E departments. Added to a shortage of some 50,000 doctors and nurses across the UK, AI could be seen as the only viable long-term solution. Virtual agents, it is argued, could handle routine enquiries, freeing up frontline hospital and primary care staff and making the whole system more efficient.
Cost savings and faster diagnosis
A report by the think tank, Reform, suggested that AI could predict which individuals or groups may be at risk of illness and either enable them to self-care or direct them to the appropriate services. It also said that diagnosis could be performed 30 times faster and more accurately than a human diagnosis, although the technology should act as an enabler rather than a decision-maker.
AI is seen as a solution to the lack of funding in the NHS too. Additional money is often ploughed into supporting a system that is, at its core, dysfunctional. A significant investment into AI, however, could enable the cleansing and digitising of patient data, which could then be inputted into the technology for accurate results. Of course, AI operating on misinformation could wreak havoc across the service and the continued security of patient data is fundamental to virtual assistance.
Moving things forward
Reform’s report identified ways in which AI could be introduced. The suggestion was to create a range of datasets that would train the agents to identify patterns or particular conditions. These could then be made available to companies wanting to produce algorithms for use in the health service. This methodology would mean funding for the research and development of the algorithms could come from business grants or R&D tax credit.
The potential gains of AI use, then, are manifold – costs could be saved long-term and there seems to be little concern over job losses, given the tens of thousands of vacancies across the service. As yet, only a few NHS trusts have implemented AI, with the system as a whole recognising the benefits but lacking clarity about where to start. Among those technologies in use is Google’s DeepMind, which identifies disease from images inside the eye.