Google Just Unveiled a New AI That Helps You Diagnose Skin Conditions

AI is transforming modern healthcare every day. AI applications are vast, spanning the scope of healthcare, from enhancing tuberculosis detection to advancing the diagnosis of breast cancer. But Google aims to use it for helping users learn about skin conditions, and whether or not they have them, according to a new preview shared in a blog post from Google's I/O team.

Notably, Google wants to roll out diagnostic AI for skin conditions on smartphones, which could allow every smartphone owner to have an idea of what their diagnosis might be. But this doesn't replace the role of a professional dermatologist.

Google's AI tool could help mitigate the shortage of dermatologists

The AI tool preview is designed to inform users of the state of their skin, nails, and hair. This is more significant than it sounds because, as the body's biggest organ, the skin is a diagnostic nexus via which health professionals can tell a lot about your general health and condition. The new AI technology employs techniques akin to those used for identifying lung cancer and diabetic eye disease via CT scans. To give possible diagnoses, the tool uses a camera to capture a skin image, and then gives possible causes for various symptoms, like rashes.

The AI is an operated dermatology technology that takes its cue from Google's data monitoring of users, who often search for skin conditions via reference photos from Google Images. And while a real-life dermatologist is still preferred, they don't work in abundance, which means diagnosing possible skin ailments for the two billion people affected by skin conditions globally per year might need some help from outsourced tools like Google's new AI tool.

Google's AI tool pulls from 65,000 skin images and case data

Deciding to leverage the flow of public curiosity online about skin ailments, Google created a new tool, and all you need to use it is a web-based application, three photos of the affected area of your skin. The app processes the images, asks users how long the symptoms have happened, and uses the answers to arrive at a more accurate list of possible diagnoses -- pulling from an archival knowledge of 288 skin conditions to give probable matches that users can investigate, or take to a dermatologist.

Every possibly matching condition will display a dermatologist-reviewed description, along with pre-written replies to frequently asked questions, and images of the skin ailment from the internet. Google, too, thinks this app is a supplement, and not a replacement for a fully-trained dermatologist, but it still hopes the AI tool can offer more organized basic information than users might find simply surfing the web. The tool will go live later in 2021, and will come pre-programmed to examine skin ailments from all races, skin tones, ages, and sexes. As of writing, the app can interpret user images via 65,000 skin images and consolidated case data of previously diagnosed conditions, including thousands of healthy skin photos and skin condition searches. After a year of stay-at-home and social distancing living conditions globally, AI has greatly expanded the capability and extended the scope of healthcare to offer users at-home assistance. Since it will take some time for everyone to achieve a level of mobility sufficient to go to the doctor, this could help to improve general health levels.