The idea of turning to Amazon.com Inc.’s gaffe-prone Alexa in a medical emergency has always seemed to be a bit of a sick joke. YouTube is packed with examples of amusing ways in which the “black, always-on cylinder the size of a Pringles can” – that’s CEO Jeff Bezos’s own description of the smart speaker – fails the Turing Test. “Alexa, I need medical assistance immediately,” says one user, before getting the soothing yet robotic answer: “I added ‘medical assistance immediately’ to your shopping list.”
That hasn’t dissuaded Britain’s National Health Service from trying out the technology for everyday health questions about common illnesses. On Wednesday it announced a partnership with Amazon to help patients get information from the NHS website via voice commands. Ideally, you should be able to ask Alexa things like “how do I treat a migraine?” and get a response sourced from the website in seconds. (Incidentally, the NHS website’s answer to that question is: “There’s currently no cure for migraines, although a number of treatments are available to help ease the symptoms.”)
The appeal to the NHS of experimenting with such tech, however superficial or unsophisticated it might seem, is a more serious matter. Britain’s free-at-point-of-use public healthcare system is under-resourced, expensive, and is set to be an even bigger strain on spending as people live longer.
The U.K.’s Institute for Fiscal Studies predicts that the British state’s healthcare bill will rise from about 154 billion pounds ($192 billion) a year to 186 billion pounds by 2023-24, and that’s in a “steady-state” scenario. This explains why the service’s managers will jump at any promise of tech-assisted efficiency gains. If people get more advice at home through automation, the theory goes, that will reduce pressure on overbooked doctors.
Yet past efforts at using automated symptom checkers haven’t found a magic fix that keeps the “worried well” from bothering the professionals. A Forbes report found that early software from London’s Babylon Health, an artificial intelligence startup that provides “health information” based on symptoms entered by users, advised people to go to the emergency room in about 30% of cases. For people dialing the NHS advice line (staffed by humans), it was 20%.
There’s also the problem of accuracy. So-called “computerized diagnostic decision support” programs aren’t up to par yet with their homo sapiens equivalents in areas like melanoma image recognition, according to a 2018 study in medical journal The Lancet.
In fairness, the new NHS-Alexa partnership seems to steer clear of direct diagnosis: It wants to be an information tool and nothing more. One public benefit would be to get more people to access official professional advice rather than make do with the rumors and misinformation that swamp social media, according to Eleonora Harwich, director of research at Reform, a public sector think tank.
Still, Amazon looks like the more obvious beneficiary of this project right now. The data might not be as rich as NHS patient records – the subject of a much more controversial partnership between the NHS and Google’s DeepMind – but they could still be lucrative in the long run.
Getting more Alexas into people’s homes, listening in to their self-volunteered symptoms, and perhaps suggesting over-the-counter treatments delivered via Amazon Prime wouldn’t be bad for Bezos’s bottom line. One demonstration of Babylon at an Amazon Web Services summit featured more than a minute of back-and-forth between human and robot before the latter recommended both speaking to a doctor and getting a prescription delivered direct to the patient’s door. Online healthcare will be serious money one day.
In the meantime, the NHS might find more savings through humdrum initiatives such as going paperless by 2020. Not as sexy as asking Alexa, but probably more useful.
To contact the author of this story: Lionel Laurent at firstname.lastname@example.org
To contact the editor responsible for this story: James Boxell at email@example.com
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Lionel Laurent is a Bloomberg Opinion columnist covering Brussels. He previously worked at Reuters and Forbes.
©2019 Bloomberg L.P.