Artificial intelligence has learned to detect diabetes by voice in 6 seconds
[ad_1]
Canadian scientists from Klick Labs have developed a tool based on artificial intelligence that can diagnose type 2 diabetes in a few seconds by voice. The creators published the study in the journal Mayo Clinic Proceedings: Digital Health.
The machine, having heard a recording of a person’s speech lasting 6-10 seconds and learned information about the state of health along with statistical data, is able to make a correct diagnosis. At this stage, AI makes the correct diagnosis in 89% and 86% of cases for women and men, respectively.
The study involved 267 volunteers, 75 of whom had confirmed disease. All participants were asked to write down the simple phrase “Hi! How are you? What is my current glucose level?” six times a day for two weeks. Next, the scientists analyzed the data array for 14 characteristics (timbre, intensity, noise level, etc.), identifying a pattern between them and the presence of a diagnosis.
The study can change the medical community’s approaches to diagnosing diabetes, Canadian scientists believe.
[ad_2]
Source link