This year, I have had a lot of personal experience with the NHS and social care crisis in the UK. It has left me with a vivid picture of the gap between our expectations of the support that is available, and the reality.
I was keenly interested in how technology and AI could help to solve the problems I was having at a personal, as well as a systematic level. Can AI help to reduce the increasing cost of healthcare?
Listening to experts in the field, I hear about AI being used in robot-assisted surgery, aiding in diagnosis, image analysis and creating new medicines. Of course, these deeply specialised activities will be of benefit to our health outcomes and hopefully the economy, too. AI is being used by some healthcare providers to detect diseases, such as cancer, more accurately and earlier. The growth of IoT and consumer wearables combined with AI is helping with lifestyle factors.
Within the care sector, I hear about use cases like virtual nursing assistants. It isn’t immediately clear to me how prevalent such virtual assistants would be, for sure most corporations use some automation in their customer engagement, and it would make sense for healthcare organisations to adopt that. Anything that streamlines the user journey in healthcare must be good.
Be part of something bigger, join the Chartered Institute for IT.
However, I worry that over-reliance on pre-scripted responses to patients would be problematic in lots of cases – my family has enough problems getting responses from the NHS without putting a chatbot in the way. Where does the decision-making require a human-in-the-loop? Or two, even? The latest compromise draft of the EU AI Act has introduced requirements for two humans to approve an AI-recommended decision rather than one – in the case of biometric identification. That’s because research has shown that human decision-making is impaired by 5–20% when approving an AI recommendation.
On the other hand, one innovation I have personally purchased is a necklace with a fall-detector and a button that does a round-robin emergency call to family members. A relatively simple innovation that can bring quite a lot of peace of mind. Another really simple use of technology was connecting a smart plug to a lamp, and then via Bluetooth to a smart speaker. This functionality restored a huge amount of independence to someone unable to reach a light switch.
There’s a lot of exciting AI work going on in the medical device sector. In November I was at WebSummit, now the biggest technology conference in the world. There was clearly a lot of innovation happening using new technologies to bring advice straight to patients as well as diagnoses. The UK market is growing rapidly, and industry is notable in driving lots of standardisation work forward. In fact, BSI Regulatory Services, one of the biggest conformity assessment providers for medical devices, is the first to announce it will apply to become a notified body under the EU AI Act.
There will be a big scramble now to align AI standards with existing medical standards, and to decide how the UK, and in particular the Medicines and Healthcare products Regulatory Agency (MHRA), will regulate and innovate in the sector. The UK has a crisis on its hands, and technology can contribute to helping solve that, but it also cannot be the only answer.