Google is taking one of the most important steps in accessing health care for a major technology company, Google, which is launching an AI-based tool that will help consumers diagnose hundreds of skin conditions.
Derm Assist is the first of its kind and will be marketed in Europe this year, before targeting nearly 2,000 million people worldwide with skin diseases ranging from acne to melanoma.
Users upload pictures of their health status through the Derm Assist website and answer questions about their symptoms. An AI model analyzes the information and creates a list of possible matching conditions. The service is free to all Internet users, whether or not they are Google users.
“The tool is not intended to provide a diagnosis. . . rather, we hope it will allow you to get authoritative information so that you can make a clearer decision about the next step, ”Google said.
It was launched after three years of development by Google. For a long time, health has been seen in the market for the benefits of advanced artificial intelligence. Apple, Amazon and Microsoft are also pushing rivals into space where they can win, building health services for consumers, medical and pharmaceutical companies.
Google chose dermatology as the primary health goal caused by AI because of the large number of people affected by skin conditions. Approximately 10 billion searches are performed each year on skin, nails and hair problems, and research has shown that only 13% of people diagnose themselves best, the huge search said.
“Skin diseases are a huge global burden as a category. People are turning to Google to investigate skin concerns. Most cases are curable, but half the world’s population is severely short of dermatologists,” said Dr. Peggy Bui, director of Google Health products and California San Francisco Internal Medicine Specialists.
The Derm Assist system is based on a machine learning algorithm trained in more than 16,000 dermatology cases worldwide. According to an examination since last year, the tool has been able to identify skin conditions as accurately as dermatologists certified by the U.S. board.
One of the information provided to users is reviewed by human dermatologists. If a user mentions alarming symptoms, such as inability to breathe, additional alerts are recommended to consult a physician immediately.
An examination published They found at the JAMA Network Open that the AI tool has significantly improved the diagnostic accuracy of non-specialists, such as doctors and nurses, to help diagnose skin conditions.
“Our observations suggest that AI has the potential to increase capacity[generalist doctors and nurses]. . . to more effectively diagnose and detect skin conditions, “wrote the author of the study by Yuan Liu and his team in a paper reviewed by classmates.” Improving the accuracy of diagnosis of unmentioned cases… can have tremendous implications for health systems.
Eric Topol, a professor of molecular medicine at the Scripps Research Institute and an expert in AI and health, said: “That was going to happen sometime, which was the main use of AI for in-depth medical studies in 2017 with some medical validations.”
To prevent cases of skin cancer from occurring through false negatives, the algorithm is designed to be prudent in making decisions. “When we designed this, we said we wanted to optimize it to achieve high sensitivity, especially for worrying or frightening conditions,” Dr. Bui said.
To address the privacy issues surrounding users ’health data, Google has said it will not use uploaded images for advertising purposes, and will only store images to train the Derm Assist algorithm if users give explicit permission to do so.
“Users control their data with the ability to store, delete, or provide data for research,” Dr. Bui said. “We hope to encourage donation, as algorithms are as good as trained data…. We will continue to improve the model by acquiring other data sets from other sources in addition to the data provided.”