AI Can Detect Type 2 Diabetes from Your Voice

Original Article:

Diabetes and the way you talk

Artificial Intelligence (AI) is becoming a powerful tool in the medical world, and now it can be used to detect signs of type 2 diabetes (T2D) just by listening to someone talk. This exciting development could make diagnosing diabetes easier and faster.

Voice Changes Linked to Health:

Our voices can change when we’re emotional, sleepy, or under the influence of alcohol. These changes can be linked to our health. Now new research suggests that AI can pick up on subtle voice changes that may indicate shifts in our well-being.

AI and type 2 diabetes

A group of researchers from Klick Labs conducted a study involving 267 participants, some with T2D and some without. Participants recorded a specific phrase several times a day for two weeks using a smartphone app. AI was used to analyze these recordings and identified 14 vocal characteristics, such as pitch and intensity. Using this data, the AI was trained to recognize the differences in voices between people with and without T2D.

Early results but this could be another success for AI

The AI was quite accurate, with an 89% success rate for women and an 86% success rate for men. Interestingly, the vocal signals indicating a diagnosis of T2D were different for men and women. Men’s voices were more affected by intensity and amplitude, while women’s voices showed variations in pitch. Although more testing is needed, this method could potentially make T2D diagnosis easier and faster.