A team of researchers at the Luxembourg Institute of Health (LIH) has unveiled an innovative AI-driven tool designed to detect type 2 diabetes through voice analysis. This groundbreaking technology aims to expedite and broaden access to diabetes screening by identifying subtle vocal changes associated with the condition.
Led by Dr. Guy Fagerazzi, the LIH team discovered that individuals with diabetes exhibit unique voice patterns, often imperceptible to the human ear. These changes, attributed to factors such as hyperglycemia, fatigue, and neuropathies, typically result in a hoarse voice, according to Dr. Fagerazzi.
Using data from the Colive Voice project, which gathered over 8,000 voice samples in multiple languages, the researchers focused on English-language recordings from 600 participants in the United States. Participants were asked to read a short passage aloud, and AI was used to analyze the recordings for specific patterns linked to type 2 diabetes.
Dr. Fagerazzi highlighted that the AI system successfully differentiated between individuals with and without diabetes, assigning distinct scores to each group. While the technology is not meant to replace traditional diagnostic methods, such as blood tests, it serves as an efficient and non-invasive preliminary screening tool for identifying individuals at risk.
With nearly 800 million people worldwide living with diabetes, half of whom remain undiagnosed, this AI tool could provide a fast and affordable means of detection. Currently, the system boasts an accuracy rate of approximately 75%, with hopes for further improvement as the technology evolves.
In conclusion, Dr. Fagerazzi emphasized that while voice analysis holds great promise, it should complement, not replace, conventional diagnostic techniques.
Related topics:
Pinwo Food’s Losses: Imported Milk and Cheese in Trouble.
