A new study has found that an artificial intelligence (AI) can accurately detect changes in a person’s voice that indicate the presence of diabetes. The technology is designed to help identify people with undiagnosed diabetes. Researchers at Klick Labs, led by Jaycee Kaufman, developed the AI, which can determine whether a person has type 2 diabetes by analyzing six to ten seconds of their voice, along with basic health data such as age, gender, height, and weight. The AI uses a range of vocal characteristics, including changes in pitch and intensity, to diagnose the disease.
To train the AI, the researchers recorded the voices of 267 people, some of whom had been diagnosed with type 2 diabetes and some of whom had not. The participants were asked to speak a sentence into their smartphone six times a day for two weeks, resulting in over 18,000 voice samples. The researchers were able to identify 14 acoustic features that differed between people with and without type 2 diabetes.
The study, published in the journal Mayo Clinic Proceedings: Digital Health, found that the AI was able to diagnose diabetes with high accuracy based on vocal characteristics. However, there was a slight difference in accuracy between men and women, with the AI achieving 89% accuracy for women and 86% accuracy for men. Yan Fossat, Vice President of Klick Labs, said that the study showed the potential of voice technology to screen many people and identify those with undiagnosed type 2 diabetes.
The researchers believe that voice technology could revolutionize medical practice as a low-cost and easily accessible digital screening tool. They plan to expand their research to investigate whether other conditions, such as high blood pressure and prediabetes, can also be detected through changes in the voice. This new technology could help to identify people with diabetes earlier, leading to earlier treatment and better health outcomes.
