Google is developing an AI-powered audio analysis tool to detect underlying diseases by identifying certain sound patterns in human voices.
Since the beginning of this year, Google has been working on Health Acoustic Representations (HeAR), an AI-powered diagnostics tool that works using sound. The bioacoustic foundation model is specially designed to listen to human sounds and flag potential early warning signs of diseases.
Google’s research team has trained the AI model on 300 million datasets of audio extracted from a diverse and de-identified group of individuals. The team has trained HeAR on over 100 million cough sounds that can help identify early signs of Tuberculosis in humans. The initiative has gained support from global welfare organizations, including the United Nations.
The United Nations has partnered with Google through its StopTB initiative, which aims to eliminate Tuberculosis from the face of the world by 2030. With an accurate early detection tool, their mission can be boosted significantly. “Solutions like HeAR will enable AI-powered acoustic analysis to break new ground in tuberculosis screening and detection, offering a potentially low-impact, accessible tool to those who need it most,” opined Zhi Zhen Qin, digital health specialist with the Stop TB Partnership.
Google has strategically teamed up with an Indian startup, Solicit Technologies, to shrink the technology into people’s mobile devices to enable real-time tracking and detection of diseases. Salcit is a respiratory healthcare AI startup that aims to help populations with poor access to healthcare and high risk of respiratory diseases.
Sniffing out diseases has become an attractive diagnostic trend ever since an old woman showcased her ability to smell Parkinson’s disease in humans by smelling them. With the deeper integration of AI in healthcare, early detection will empower the success rate of preventing and curing critical illnesses that become incurable after a certain growth stage.
