Most people are aware of biometrics used to protect your privacy or give special access. Though in these cases we think first of retina scans, fingerprints, or face scanners. However, an article by NPR describes the novel idea of using voices to detect illness, though it is still in its development stages (Weiner, 2022). The study is being carried out by researchers in over 10 institutions, with the goal of amassing a lot of sample voices to create a big database where they can study human voices with AI.
The question driving it is ‘If one could only listen to a patient when they come in with a problem, would the physician be able to form a diagnosis?’ Obviously, it is easy to guess that a person with lung problems would sound phlegm-ridden, but other diseases are not easily imaginable. The article continues by stating that Parkinson’s disease or cancer patients also have a key pace of voice, and there are even accompanying voice recordings (but I must warn that they are a bit repulsive).
Surprisingly, there are actual companies that implemented this idea, one such is audEERING, a Germany-based company that uses AI to perform voice analysis. Their applications are used in game testing and the health industry to detect emotions but also health information (audEERING, 2022). The former can be used to gather data on the emotions or reactions of players through their voices. Though the latter is focused mainly on CoVid-19 patients. Their applications are thus not that advanced to be able to diagnose all diseases, in contrast to the study which is part of NIH’s Bridge to AI program.
It seems that the research for voice analytics could turn out to be useful, the NPR article concludes by stating that it could ultimately be integrated with our phones or other communication devices and that they could alert us when they detect an illness. More so with diseases like Parkinson where early detection is key for receiving timely treatment. Even though details like voice recording ownership for the samples and commercialization of the technology are hindrances that need to be figured out. Voice Analytics seems like a great innovation to improve how we get diagnosed.
Sources:
audEERING. (2022, September 28). Company. audEERING. Retrieved October 16, 2022, from https://www.audeering.com/company/
Weiner, L. (2022, October 10). Artificial intelligence could soon diagnose illness based on the sound of your voice. NPR.Org. Retrieved October 16, 2022, from https://www.npr.org/2022/10/10/1127181418/ai-app-voice-diagnose-disease
Hi Shadiah, thank you for sharing this thrilling application of AI. I agree that it has great potential for diagnosing diseases at an early stage, making it possible to give patients timely treatment, but also just as you said about the underlying user privacy issue, I believe the requirement of constant voice recording might be one of the biggest impediments of the commercialization of the technology. Do you think people will be willing to ‘compromise’ their privacy with those huge benefits that can be brought by the advanced technologies in the future? I’d say probably yes, especially when it comes to the benefits associated with human lives. So as far as I am concerned, the challenge that matters would be how are we going to address the data privacy issue with technologies, so that we feel safe to take advantage of them.