AI System Screened for Diabetes Using Vocal Patterns
Researchers developed a tool that detects type 2 diabetes by analyzing vocal changes from short audio recordings.
Updated on Sept. 28, 2026 in Diabetes
Live Poll
Would you trust an AI tool to help screen for your potential health conditions?
Engineers from thymia and RMIT University have created an artificial intelligence model capable of identifying type 2 diabetes through vocal analysis. The system successfully flagged 80 percent of UK study participants who reported living with the condition.
Why it matters
This technology aims to reach undiagnosed individuals who do not undergo routine health screenings, potentially expanding early detection beyond traditional blood testing. It offers a non-invasive alternative for identifying risk using simple audio samples.
The model was trained on 63,000 voice recordings from 21,000 subjects in the UK and US. A sub-group analysis of 801 participants showed a 75 percent accuracy rate when compared against results from home blood tests.
The players
thymia
This health technology company specializes in developing AI-driven tools for diagnostic screening.
RMIT University
Based in Melbourne, Australia, this public research university led the development of the vocal analysis tool.
The details
The AI analyzes 20-second clips of individuals reading Aesop's fables, specifically looking for indicators like raspy speech and reduced respiratory control. Data can be collected via standard phone calls or mobile applications.
Timeline
September 28, 2026: The research findings were published.
Health Landscape
This development represents a departure from standard blood-test diagnostic protocols for individuals aged 40 to 74 by enabling remote screening via voice. It reflects a broader shift toward integrating passive, non-invasive digital biomarkers into early-stage diagnostic pipelines.
This technology could eventually allow individuals to screen for diabetes from home using only a phone call or mobile app. It may reduce the barrier to entry for those currently avoiding traditional clinical blood tests.
The takeaway
Vocal analysis presents a scalable, non-invasive method for identifying health risks before symptoms become severe. Future clinical integration could significantly increase the number of people screened for chronic conditions in non-traditional settings.
Further reading
For additional context on screening technologies, visit the Diabetes section.
Live Poll
Would you trust an AI tool to help screen for your potential health conditions?
