AI 'Speech Clock' Suggests Your Voice Could Reveal Accelerated Aging and Dementia Risk

A new artificial intelligence tool can analyze how people speak to estimate their chronological age and flag voices that sound "older" than expected—an indicator that may be linked to cognitive impairment, dementia and accelerated biological aging, a study published in Science Advances reports.

What The Study Did

Researchers drew on data from the ReD-Lat consortium, a large dementia research collaboration in Latin America. The analysis included recordings and transcripts from more than 2,900 Spanish-speaking adults (ages 18–88) across Argentina, Chile, Colombia, Mexico and Peru. About 1,500 participants were cognitively healthy; others had mild cognitive impairment, Alzheimer's disease, or frontotemporal dementia.

How The AI Works

The team extracted over 700 audio and language features—such as pauses, speech rate, pitch, vocabulary variety and emotional tone—from seven short speech tasks (for example, describing an animated clip, rapid naming exercises, and immediate and delayed story recall). A machine-learning model was trained to predict chronological age from those features. The difference between a participant's AI-predicted age and their actual age was called the "speech-age gap."

Main Findings

Participants whose voices sounded older than their chronological age (larger speech-age gaps) were more likely to:

• Have cognitive impairment or clinical dementia, with the largest gaps in language-dominant frontotemporal dementia.
• Perform worse on tests of memory, language and attention and report more difficulty with everyday functioning.
• Show biological signs of accelerated aging: larger speech-age gaps correlated with higher age acceleration on three epigenetic clocks.
• Have higher levels of p-tau217 among those with Alzheimer's disease, a protein linked to Alzheimer's pathology.

Social and Health Correlates

The study also found that larger speech-age gaps were associated with socioeconomic disadvantages—financial strain, food insecurity, limited access to healthcare, difficult childhood experiences, and lower educational attainment. These relationships were broadly consistent across the five countries studied.

Expert Perspective and Caveats

"Existing aging clocks are very powerful, but many of them rely on brain scans or blood tests that are expensive, invasive or difficult to repeat regularly. All of these limitations can be overcome with speech," said study co-author Adolfo García of the University of San Andrés.

At the same time, clinicians caution that voice changes are sensitive to many factors. "The human voice is incredibly sensitive to our overall well-being," said Dr. Manisha Parulekar of Hackensack University Medical Center. Depression, fatigue or acute stress can make speech sound older, so the tool should not be used alone to diagnose dementia.

Limitations And Next Steps

The study was largely cross-sectional: many participants were assessed at a single time point, so researchers could not determine whether an older-sounding voice predicts future cognitive decline or instead reflects changes that accompany existing decline. The model was trained and tested only on Spanish-speaking participants from Latin America, so language-specific adaptation and validation are required before applying the method in other languages or regions.

Bottom Line

The "speech clock" is a promising, low-cost, noninvasive tool that could help flag people for closer cognitive evaluation and research. However, it requires longitudinal validation and translation across languages and populations before it can be used clinically. This article is for informational purposes only and does not provide medical advice.