Google has developed an AI model that analyzes sound signals to detect early signs of disease, Bloomberg reports.

This technology can work on smartphones and help in places where it is difficult to access medical equipment, such as X-ray machines. The phone's microphone can replace expensive diagnostic equipment and help in diagnosis.

This AI system is already helping to fight tuberculosis, which kills almost 250,000 people in India every year. Although the disease is treatable, millions remain undiagnosed. Early detection is key to stopping its spread.

Google trained its model on 300 million sound fragments, including coughs, sneezes, and breathing. This data was collected from open sources, such as YouTube videos, as well as recordings of coughs of patients in a hospital in Zambia. Body sounds contain information about our health, which helps diagnose diseases.

One Indian company, Salcit Technologies, has integrated this model into its app to improve the accuracy of tuberculosis diagnostics and lung health assessments. Their app, called Swaasa, allows users to record a 10-second cough sample and get results with 94% accuracy.

However, implementing new technologies into medical practice faces challenges. There is a need to find acceptance among doctors and ensure the app is used correctly in rural settings where users may not have sufficient experience with technology.

Google is also working on another project related to bioacoustics, using ultrasound for early detection of breast cancer in Taiwan. AI helps detect tumors, and Google plans to roll out this technology globally.

While none of Google’s models are ready for commercialization yet, sound-based AI could make early diagnosis more accessible and affordable.