Will Google’s AI be able to detect diseases with a cough?
Google, together with the Indian startup Salcit Technologies, is creating a new artificial intelligence technology capable of diagnosing tuberculosis through a voice recording. This could facilitate the diagnosis of other respiratory diseases as well. All the details
Artificial intelligence (AI), just as it is available today, is capable of identifying the signals launched by certain diseases early and therefore its development can take us even further.
Google, for example, is working with the Indian startup Salcit Technologies, which specializes in artificial intelligence (AI) for respiratory care, on a new technology capable of diagnosing tuberculosis using voice recordings from smartphones.
DECRYPTING OUR BODY'S SIGNALS
From a cough to breathing, the sounds our bodies make are full of information about our health. “The subtle clues hidden in these bioacoustic sounds have the potential to revolutionize the way we test, diagnose, monitor and manage a wide range of health conditions such as tuberculosis (TB) or chronic obstructive pulmonary disease (COPD),” he says Google.
According to big G, decrypting these sounds can help intervene early even in places where healthcare facilities do not always have advanced equipment. And all this thanks to simple and widely accessible smartphone microphones. Artificial intelligence is then useful for extracting health information from acoustic data.
GOOGLE'S AI TO DIAGNOSE DISEASES
The Mountain View giant is in fact training one of its AI models to listen, via smartphone microphones, to the signals of certain diseases using sounds such as coughing and sneezing. And earlier this year it introduced Health Acoustic Representations (HeAR), a bioacoustic baseline model, trained on 300 million audio data points, including 100 million cough sounds, to learn to spot patterns in sounds.
COLLABORATION WITH SALCIT FOR TUBERCULOSIS
Google is also collaborating with Salcit Technologies, which is testing HeAR in combination with its product Swaasa, which uses AI to analyze cough sounds and evaluate the health of the lungs. The goal is to try to improve early diagnosis of tuberculosis based solely on cough sounds.
Tuberculosis is a treatable disease, but millions of cases go undiagnosed every year, often because people lack access to health services. In fact, according to the United Nations-backed nonprofit The Stop TB Partnership , between 3 and 4 million cases go unreported and, if left untreated, the disease's mortality rate exceeds 50%.
THE POTENTIAL OF AI IN MEDICINE
The rapid development of artificial intelligence is producing new opportunities for the early diagnosis of a wide range of diseases: from identifying the signs of chronic diseases, such as diabetes and hypertension , to those of some types of endometrial cancer , to to the early identification of Parkinson's disease . Additionally, the University of California, Los Angeles (UCLA) recently said it is developing a new AI-enhanced test that could help speed up the process of diagnosing Lyme disease.
As Axios explains, many of today's diagnoses are limited by whether there is a known biological marker for a disease or whether the doctor knows exactly what to look for, but AI stands out and can make a difference in research because of its ability to use and analyze a huge amount of data.
And, finally, it could also prove useful for identifying diseases that the doctor is not looking for because, as the CEO of the AI database startup Dandelion Health, Elliott Green, highlighted: “Artificial intelligence goes where there is evidence, not where we think the evidence should be.”
This is a machine translation from Italian language of a post published on Start Magazine at the URL https://www.startmag.it/sanita/ia-intelligenza-artificiale-google-malattie-con-un-colpo-di-tosse/ on Tue, 03 Sep 2024 12:08:15 +0000.