AI Tool Has Screened for Type 2 Diabetes via Voice
Researchers developed a system that uses vocal patterns to identify individuals at risk for type 2 diabetes.
Updated on Sept. 28, 2026 in Diabetes
Live Poll
Would you trust an AI tool to help screen for your potential health conditions?
Engineers have developed an artificial intelligence tool that identifies type 2 diabetes by analyzing short vocal recordings. This technology, which screens for indicators like raspy speech, could help identify undiagnosed cases in populations that skip routine blood screenings.
Why it matters
Type 2 diabetes often goes undetected in people who do not attend routine medical checkups. This non-invasive tool aims to improve early detection by making preliminary risk assessment more accessible.
In an evaluation of 7,319 UK participants, an AI model trained on over 63,000 voice recordings from 21,000 subjects identified 80 percent of self-reported diabetes cases. A sub-group analysis of 801 participants found the algorithm correctly flagged 75 percent of cases confirmed by blood tests.
The players
RMIT University
A public research university in Melbourne, Australia, that focuses on engineering and technology solutions.
thymia
A technology company specializing in the use of artificial intelligence to assess health through behavioral and vocal analysis.
The details
The model detects subtle vocal changes that may accompany the disease, including increased raspiness and impaired respiratory control. Users read 20-second clips of Aesop's fables, with the AI analyzing these audio samples collected via phone calls or mobile applications. By focusing on these acoustic biomarkers, the system attempts to capture physiological shifts associated with diabetes without requiring a blood draw.
Timeline
September 28, 2026: The research findings were published.
Health Landscape
Current diagnostic standards for type 2 diabetes rely heavily on blood tests for adults aged 40 to 74. This vocal analysis approach marks a potential shift toward non-invasive, digital screening methods designed to reach individuals outside of traditional primary care settings.
This tool is an experimental screening method and does not replace established blood-based testing. If you are concerned about diabetes risk, it is worth discussing appropriate screening schedules with your doctor.
The takeaway
Vocal analysis may eventually serve as a low-barrier screening tool for metabolic health. If you are age 40 or older, prioritize regular blood-based diabetes screenings as recommended by your physician.
What happens next
The research team plans to conduct future evaluations of the vocal analysis model within clinical settings.
Further reading
For more on evolving diagnostic technology, visit the Diabetes section.
Source note: This article includes information reported by The Independent.
Live Poll
Would you trust an AI tool to help screen for your potential health conditions?
