How artificial will AI in healthcare be?

If we mention ‘technology’ and ‘healthcare’ what things come to mind? A big magnetic resonance imaging (MRI) scanner? An ambulance maybe. A heart monitoring machine, by the side of a bed.

Technological progress in healthcare has until now (broadly speaking) mostly been about giving medical professionals tools to make us, the patients, better. While in the main it’s been good for all of us, it hasn’t changed the fundamental nature of power in the doctor-patient relationship.

The combination – it’s actually truer to say ‘combinations’ – of mobile technology, the internet, big data and artificial intelligence (AI) are poised to change that. We’ll soon be entering a landscape where patients are empowered to take control.

Many of us have already got used to using Google and other online sources of information to try and find out what’s wrong with us, or get consultations at a time of our choosing. Meanwhile, our smartphones allow us to scan barcodes so we can make healthier eating choices, provide us with support if we’re trying to stop smoking (got a craving? Press a button for distraction!) – even diagnose everything from skin cancer to Parkinson’s disease.

But it is the potential – the rapidly-arriving reality – of what AI will bring to the patient experience that is really interesting. No doubt we’ve all read the headlines about IBM’s Watson being able to diagnose cancer, and Alphabet’s DeepMind using machine learning to spot injuries to kidneys.

And when AI and mobile technology come together, we expect to see a slew of apps that will really help patients with different conditions directly. Take, for example, Woebot (woebot.io), a cute animated robot that asks you questions about how you are feeling and what your energy levels are like. In reality it’s a chatbot using AI to deploy cognitive-behavioural therapy techniques to help people deal with depression and anxiety.

Similarly, Ada (ada.com) is an app that uses AI and real-life medical cases to train itself so it can help people self-diagnose. After users describe their condition, Ada asks a series of questions about their possible symptoms, cross-referencing answers against self-provided medical information. People get a summary of likely conditions, with the option to chat online with a physician for extra peace of mind.

As patients stream medical data from smartphones and wearables, this positive feedback loop will only grow – and so healthcare AIs will learn to do more and more. Future apps and programmes could, for one, provide automatic diagnosis from a description of symptoms and use Instagram to spot behavioural traits that suggest you are depressed.

The underlying idea behind these advances is clear; hopefully you as a patient ‘feel good’ (or at least better) by playing an active role in managing your health; and by agreeing to share your data, you are also ‘doing good’ by helping to train medical algorithms for the future.

Is there a limit to the empowerment of patients that AI can provide? We may get diagnostic facts, but will we get empathy, understanding and maybe even humour? There’s an inherent human element to healthcare that however much technology moves forward will always be needed. Artificial intelligence is fine; but artificial compassion will never catch on.