Google – in its latest effort to convince the public that is actually isn’t a sinister company that knows more about you than your own mother – has revealed a new AI-powered tool to spot skin, hair, and nail conditions.
The tool analyses a photo of the area of concern and cross references it with Google’s library of medical images. AI systems have been built around this concept for the last three years and will be ready for launch later in 2021.
The system can currently recognise nearly 300 skin conditions with a high degree of accuracy, but Google is quick to insist that this is not a replacement for diagnosis, but rather another method of increasing efficiencies in the patient triage space.
Ideally, the tool will allow patients to be more precise when describing their conditions, while healthcare professionals will gain another tool in their diagnosis arsenal.
The real test of the tool’s usefulness may come from our willingness to trust AI over a human’s opinion.
AI-based diagnosis systems are already much faster – and, in a lot of cases, more reliably accurate – than their human counterparts, imagine what the technology will be like in ten years? For humans living now, however, the second opinion of a human being with abstract reasoning and emotional sympathy is somewhat comforting.
The real application of this brand of technology will come in its decentralising of the triage process. Almost all patients now own a smartphone, and the majority of these will be capable of running AI-based medical software like this.
With many of us wearing Garmin, or Apple watches our phones are already analysing heart rate and blood information such as O2 saturation. In future many other key vitals will all be monitored just by using the hardware available on patients’ wrists and phones.
This – I hope – will spare others the situation I was privy to whereby by doctor had to facetime a video of my bare chest to their dermatologist friend to find out what exactly was wrong with me; the cold and efficient calculating of an AI-powered app suddenly sounds a lot more appealing.
The AI can recognise 288 skin conditions but is not designed to be a substitute for medical diagnosis and treatment, the firm said.
