Google's latest AI tool claims to identify common skin conditions

The web app, previewed today at I/O, could launch later this year.


During its opening-day Google I/O keynote today, the company announced a new AI-powered "dermatology assist" tool that could help anyone with a smartphone get more information about common skin conditions. According to Google, the project as been in the works for three years and it hopes to launch it a pilot program later this year.

The tool is actually a web-based application that you use along with your smartphone's camera. After taking three photos of your skin, hair or nails from different angles, you'll be asked to answer questions about your skin type, the issue you're currently experiencing and any symptoms you may have. The AI model, which is built upon knowledge of 288 conditions, then analyzes the information you provided and gives you a list of possible matches.

The results will be built upon dermatologist-reviewed information, answers to frequently asked questions and similar images from search results. According to Google, the model already takes into account age, sex, race, skin types and other factors that could influence results.

Considering the hair, skin and nails make up the largest organ of the body, it's understandable why Google would expand its health-monitoring efforts with a tool like this. However, like most other consumer-facing health tech, the tool isn't meant to replace a visit to your doctor or serve up diagnoses Rather, it's only designed to give you more "authoritative information" based on the photos you provide and the more detailed questions you'll answer.

In addition, Google announced a separate AI-powered tool that helps identify potential tuberculosis (TB) patients for follow-up testing. Built upon the company's existing work in medical imaging, the screening tool uses a deep learning system that can identify possible TB patients based on their chest X-rays. As with the dermatology assist tool, Google built the TB model on de-identified data from nine countries to take into account a wide range of races and ethnicities. Using a tool like this could help "save up to 80 percent of the cost per positive TB case detected."

Google will continue developing its TB screening tool later this year with two new research studies. As for the dermatology assist tool, those interested in gaining early access can sign up here.