• AiNews.com
  • Posts
  • Google Develops AI to Detect Sickness Using Sound Signals

Google Develops AI to Detect Sickness Using Sound Signals

Author

Alicia Shapiro
August 30, 2024 • Estimated Reading Time: 3 minutes

An image showing a smartphone with a visual representation of sound waves emanating from it, symbolizing the AI’s ability to detect signs of sickness through audio. The screen displays the HeAR AI logo along with icons representing health (such as a stethoscope and heartbeat). In the background, there are subtle representations of medical settings, like a hospital and respiratory symbols, indicating the focus on healthcare. The image conveys the innovative concept of using AI and sound to diagnose illnesses

Image Source: ChatGPT-4o

Google Develops AI to Detect Sickness Using Sound Signals

Google is advancing its AI capabilities by exploring the use of sound signals to predict early signs of disease. The company has trained an AI model using millions of audio samples to identify subtle indicators of illness, such as coughs and labored breathing, and is now collaborating with partners to bring this technology to high-risk populations.

HeAR: Google’s Bioacoustics-Based AI Model

Google's AI model, known as HeAR (Health Acoustic Representations), leverages the field of bioacoustics—a blend of biology and sound—to gain insights into how certain sounds can reveal early signs of sickness. HeAR was trained on 300 million two-second audio samples, including coughs, sniffles, sneezes, and breathing patterns, collected from publicly available content on platforms like YouTube.

According to Bloomberg, the field of bioacoustics provides subtle, almost imperceptible clues that can indicate early signs of illness, offering valuable insights for healthcare professionals in diagnosing patients. Additionally, the AI model can identify tiny variations in a patient’s cough patterns, which could help detect the early onset or progression of a disease.

Detecting Tuberculosis and Other Respiratory Illnesses

One of the key applications of HeAR is in detecting tuberculosis. The AI model has been trained on 100 million cough sounds, enabling it to identify signs of the disease with high accuracy. In regions where access to quality healthcare is limited, this technology could offer an alternative diagnostic tool using just a smartphone's microphone.

Collaboration with Salcit Technologies

To enhance the effectiveness of HeAR, Google has partnered with Salcit Technologies, an AI healthcare startup based in India. Salcit’s own AI model, Swaasa (meaning "breath" in Sanskrit), is being used to improve HeAR’s accuracy for tuberculosis and lung health screening. Swaasa offers a mobile app that allows users to submit a 10-second cough sample, which can detect diseases with an accuracy rate of 94%.

Affordable and Accessible Healthcare

The auditory-based test provided by Swaasa costs just $2.40, significantly cheaper than traditional spirometry tests, which can cost around $35 in Indian clinics. This affordability makes the technology accessible to populations in regions with limited healthcare resources.

Challenges and Future Prospects

Despite its potential, HeAR is still in the early stages of development and faces challenges, such as dealing with background noise in audio samples.

While Google’s bioacoustics-based AI model is still far from being market-ready, the innovative use of AI combined with sound in the medical field shows great promise. This approach could revolutionize how we diagnose and monitor health conditions, offering a new frontier in healthcare technology.