AI 'speech clock' estimates aging speed from your voice

Researchers unveiled an AI tool that estimates a person's age from hundreds of speech and language patterns. In a study of nearly 3,000 Spanish-speaking adults, a larger gap between estimated and actual age correlated with cognitive impairment, dementia and accelerated biological aging.

• Model trained on recordings of 2,900 people aged 18–88 across five Latin American countries
• Over 700 speech features analyzed, including pauses, pace, pitch, vocabulary and emotion
• Largest speech-age gaps seen in language-dominant frontotemporal dementia
• Bigger gaps linked to faster aging on three epigenetic clocks and higher p-tau217