India Utilises AI to Detect Tuberculosis

A team of researchers in India has begun employing a Google AI model to detect tuberculosis by analysing coughs. This is detailed in a report by Google Research.

In March, Google launched Health Acoustic Representations (HeAR) — a bioacoustic model for analysing sounds and detecting early signs of diseases. It is trained on 300 million audio data, including 100 million cough sounds.

Based on HeAR, the Indian company Salcit Technologies developed a product called Swaasa, which uses AI to assess lung health. The current focus is on detecting tuberculosis. The firm is exploring the potential for expanding analytical capabilities using artificial intelligence.

“Tuberculosis is a treatable disease, yet millions of cases go undiagnosed each year, often due to lack of convenient access to medical services. Improving diagnostics is crucial for eradicating tuberculosis, and AI can play a significant role in its detection, making care more accessible and affordable for people worldwide,” the report states.

Previously, scientists taught the neural network EMethylNET to detect cancer with 98.2% accuracy based on DNA data from tissue samples.

Earlier, experts at MIT developed an AI model for analysing medical images.