Introducing an AI that can tell whether you have type 2 diabetes in a few seconds based on the way you speak

Klick Labs, a research institute of Canadian medical startup Klick Health, has developed a method to highly accurately identify type 2 diabetes by combining AI and voice technology. The disease detection accuracy of this model is 89% for women and 86% for men, and the paper is published in ``Mayo Clinic Proceedings: Digital Health''.

Acoustic Analysis and Prediction of Type 2 Diabetes Mellitus Using Smartphone-Recorded Voice Segments - Mayo Clinic Proceedings: Digital Health

AI and 10 seconds of voice can screen for diabetes, new study in Mayo Clinic journal reveals | Klick Health®

The Way You Speak Can Reveal Whether You Have Type 2 Diabetes : ScienceAlert
https://www.sciencealert.com/the-way-you-speak-can-reveal-whether-you-have-type-2-diabetes

Researchers at Klick Labs asked 267 people, both diagnosed with type 2 diabetes and those without diabetes, to record certain phrases on their smartphones six times a day for two weeks. , a total of 18,465 audio data were obtained.

The researchers processed this audio data and found differences in 14 acoustic features between non-diabetic and type 2 diabetic patients, such as changes in pitch and intensity that are imperceptible to the human ear.

When AI was trained by labeling voice data with factors such as gender, age, BMI, and presence of type 2 diabetes, the AI was able to determine whether a person had type 2 diabetes with an accuracy of 89% for women and 86% for men. I was able to. Currently, to diagnose type 2 diabetes, it is necessary to draw blood and analyze it, but with a new method, detection can be made by simply recording audio with a smartphone.

Researchers say testing with larger and more diverse groups of people is needed to verify this result.

According to Jaycee Kaufman, lead author of the paper, men and women had different acoustic characteristics, with changes in intensity and amplitude being more important for men, and changes in pitch being more important for women.

'Our study reveals significant differences in vocalizations between people with type 2 diabetes and those without, and has the potential to revolutionize the way the medical community screens for diabetes,' Kaufman said. There is.”