Indonesian Political, Business & Finance News

AI Can Now Analyse Voice to Detect Signs of Disease

| | Source: Media Indonesia Translated from Indonesian | Technology
AI Can Now Analyse Voice to Detect Signs of Disease
Image: Media Indonesia

The human voice is not only a tool for communication but can also store information about a person’s health.

Advances in artificial intelligence (AI) now allow researchers to analyse small changes in how someone speaks, breathes, and the quality of their voice to look for signs of disease.

The technology, known as vocal biomarkers, has the potential to help detect and monitor various conditions, including Parkinson’s, Alzheimer’s, depression, heart failure, and type 2 diabetes.

Recent research shows that changes in voice can provide clues about physiological and cognitive processes occurring in the body. AI can be used to process very subtle voice characteristics, so changes that are difficult for humans to recognise can be analysed as health data.

According to the University of South Florida (USF Health), researchers and clinicians from the USF Morsani College of Medicine and the Luxembourg Institute of Health are developing a scientific framework for vocal biomarkers.

The effort involves 24 international experts from Europe and North America and is part of the VOCAL initiative, which aims to establish standards for voice biomarker research.

The development of voice biomarker technology is supported by the National Institutes of Health (NIH) through the Bridge2AI programme, launched in 2022 with an investment of up to US$130 million over four years to expand the use of AI in biomedical and behavioural research.

One of its projects, Bridge2AI-Voice, focuses on ethically collecting voice, speech, and cough data linked to health information, so that researchers can use it to develop and test AI models in recognising voice patterns potentially associated with various diseases.

The potential of this technology is seen in a Bridge2AI-Voice study that tested an AI model for screening laryngeal lesions, including benign, precancerous, and cancerous conditions, using data from 205 participants and 131 acoustic features from voice recordings.

The results showed that voice analysis can help distinguish participants with laryngeal lesions from a control group, although the technology still requires further testing and validation before it can be widely applied in healthcare.

One advantage of voice biomarkers is that data collection is relatively simple and non-invasive because it can be done through digital devices, potentially supporting remote health monitoring and early screening.

However, a Bridge2AI-Voice feasibility study shows there are still challenges in implementation. Of 47 participants who had smart devices, only 41 per cent successfully completed the acoustic recording task, indicating that the technology still needs improvement, particularly in terms of app design and ease of use.

Although promising, AI cannot yet be used to diagnose disease based on voice alone because vocal biomarkers still need to be tested to ensure they are consistent, accurate, and applicable across various populations.

USF and the Luxembourg Institute of Health consider uniform scientific standards necessary, which is why the VOCAL framework was developed to distinguish voice measurements from validated biomarkers and to support the development of validation and regulatory standards.

If further developed, this technology has the potential to support non-invasive disease screening and monitoring, but for now AI analysis results remain a supporting tool and do not replace a doctor’s diagnosis.

More from 7 August 2026 →

Cite this article
Source: jawawa.id (https://jawawa.id/newsitem/ai-can-now-analyse-voice-to-detect-signs-of-disease-1786121488)
Originally published by Media Indonesia (https://mediaindonesia.com/teknologi/919561/ai-kini-bisa-menganalisis-suara-untuk-mendeteksi-tanda-tanda-penyakit), translated from Indonesian
Print