We select and review products independently. When you purchase through our links we may earn a commission. Learn more.

Deep-Learning AI is Helping Google Assess Health Issues Directly in Search

Graphic of person using Google's AI tool to assess skin rash

At its most recent I/O event, Google announced it has a new deep-learning AI tool that can help detect, identify, and assess health issues, like skin conditions or tuberculosis (TB), more efficiently. And you can use it on your smartphone.

The tools use your device’s camera in tandem with Google’s AI tech. The technology is designed to help users become more educated about their health, and to stay more informed and take educated and guided steps towards a remedy.

The Dermatology Assist Tool

Google’s AI-powered dermatology assist tool makes it easier for you to better understand common issues with your skin, nails, and hair. It uses many of the same techniques used to detect diabetic eye disease or lung cancer in CT scans. And Google is using it to help you get answers about, say, a rash or weird spots on your skin.

Google serves up answers for more than 10 million skin-, nail- and hair-related issues each year, which proves that most people start looking for answers online before heading to a doctor. This tool, then, takes the form of a web-based application that is set to launch later this year.

Once launched, you’ll use your device’s camera to take three photos of the area in question from three different angles. From there, you will answer a short series of questions about your skin type and how long you’ve had the issue or symptoms for. Google’s AI model will analyze that information and run it against its database of 288 conditions, which will allow it to pull up a list of possible matching conditions. 

For each matching condition Google returns, the tool will show you dermatologist-reviewed info along with commonly asked questions and similar matching images. From there, you can further your research or make the decision to visit a doctor on your own. The tool is not intended to replace your doctor, an in-person examination, or testing; rather, Google says “we hope it gives you access to authoritative information so you can make a more informed decision about your next step.”

Using the Tool to Improve Tuberculosis Screenings

Google's AI tech helping screen tuberculosis x-rays

In addition to its dermatology assist tool, Google also shared research about how it is using its AI-based screening tool to help “identify potential tuberculosis patients for follow-up testing.” Google is also contributing to the World Health Organization’s “The End TB Strategy” to help reduce instances of the disease.

Currently, TB affects roughly 10 million people every year, and disproportionately infects those in low-to-middle income countries. Early detection is key, but it’s still quite difficult as its symptoms are much the same as those from other common respiratory diseases. And while cost-effective screening (like chest X-rays) help, experts aren’t always around to interpret the results. Google’s AI tool can help change that, saving time and money along the way.

The company’s deep-learning system can successfully and accurately identify patients who most likely have active pulmonary tuberculosis from an x-ray. The screening tool will be implemented within the process as a step before a more expensive diagnostic test is ordered. This can potentially save patients 80% of the cost per positive TV case.

The tool has a false-negative and false-positive rate similar to 14 radiologists, even in patients with HIV (which makes it more difficult to detect). Google also tested the tool on de-identified data from patients across five countries, to help it more accurately work for a wider variety of races and ethnicities.

To apply these findings in the real world, Google calibrated the thresholds of the AI system, which produces a number between 0 and 1 as a TB risk indicator. The research “suggests that any clinic could start from this default threshold and be confident that the model will perform similarly to radiologists, making it easier to deploy this technology. From there, clinics can adjust the threshold based on local needs and resources.”

With global efforts underway, the World Health Organization hopes this—along with earlier screenings—helps reduce the number of future cases in the next decade.

Source: Google

Suzanne Humphries Suzanne Humphries
Suzanne Humphries was a Commerce Editor for Review Geek. She has over seven years of experience across multiple publications researching and testing products, as well as writing and editing news, reviews, and how-to articles covering software, hardware, entertainment, networking, electronics, gaming, apps, security, finance, and small business. Read Full Bio »