AI that can “hear” signs of illness!?? Google is working on it.

Google has begun training one of its basic AI models using audio signals to predict early signs of illness.

Google’s AI division is reportedly tapping into “bioacoustics” — a field that blends biology and sound, which in part helps researchers gain insights into how the presence of pathogens affects the human voice. As it turns out, our voices convey clear information about our well-being.

According to a report by Bloomberg, the search engine giant has built an AI model that uses audio signals “to predict early signs of illness.” In places where access to quality healthcare is difficult, this technology could step in as an alternative, as users need nothing more than their smartphone microphone.

How does Google’s bioacoustics AI work?

Google’s bioacoustics AI model is called HeAR (Heath Acoustic Representations). It was trained on 300 million two-second audio samples of coughs, colds, sneezes, and breathing patterns, 100 million of which were cough sounds.

The audio clips were pulled from publicly available, non-copyrighted content on platforms like YouTube.

According to Bloomberg, bioacoustics could provide “almost imperceptible clues” that could reveal subtle signs of illness that could help health professionals diagnose patients. Additionally, the AI ​​model can detect subtle differences in patients’ cough patterns, allowing it to spot early signs of illness improvement or deterioration.

Google is partnering with Salcit Technologies, an AI healthcare startup based in India. Salcit has its own AI model called Swaasa (which means “breath” in Sanskrit), and the Indian model is using Swaasa to help HeAR improve its accuracy for screening tuberculosis and lung health.

HeAR isn’t without its challenges. For example, Google and Salcit are still working to overcome issues with users sending voice samples with too much background noise.

Google’s bioacoustics-based AI model is still a long way from being released to the market.

Source: Mashable - https://ar.rt.com/y7zo - Published on september 5-2024