Artificial intelligence-powered tools are poised to take the medical field by storm, all by harnessing a person’s voice. According to NPR, the National Institutes of Health is funding a new project that will collect voice data. The purpose of this is to teach the AI program to track, monitor, and recognize illness for diagnoses simply by on the patient’s speech patterns. For many, using the human voice as a means to monitor health and diagnose diseases sounds like a thing of science fiction.
But according to Laryngologist Dr. Yael Bensoussan, the Director of the University of South Florida’s Health Voice Center, and lead in the study, the human voice holds a great deal of information about your health. She said of the study, “We asked experts: Well, if you close your eyes when a patient comes in, just by listening to their voice, can you have an idea of the diagnosis they have…And that’s where we got all our information.”
For example, human pain normally can be heard in how people speak. Also, cases of strokes can be picked up by the slurring of words, but that’s not all. In the case of someone with Parkinson’s Disease, a slow and low speech pattern can be a key indicator of the illness. Scientists want to go further and even be able to diagnose cases of cancer and depression by using a patient’s speech patterns.
The way this will work is for the team to start with five areas of focus in collecting voice data: voice disorders, neurological disorders, mood disorders, pediatric disorders (Autism & Speech delays), and respiratory disorders. This will create the training data for the AI program. From there, it will begin to learn patterns and identify diseases over time. The project is a part of the NIH’s Bridge to AI program which launched last year. So far, it has over $100 million in funding and is tasked with creating a large-scale database to be used in healthcare for precision medicine.
The importance of data has only grown in medicine and researchers are attempting to close the gap between data and medicine. AI has only become a more important tool in most industries, but to scale and improve performance infrastructure must be created, and Dr. Yael Bensoussan understands this. “We were really lacking large what we call open source databases…Every institution kind of has their own database of data. But to create these networks and these infrastructures was really important to then allow researchers from other generations to use this data.”
But data is also the problem. Unlike other fields, medicine must contend with laws and regulations that protect patients first. HIPPA and other laws at the state level that regulate medical privacy don’t provide a clear picture if voice data that can be shared between researchers. This is even more important as it’s still unknown if the identity of the patient and their health data can be separated enough to allow for widespread data sharing and use that would be required to power an AI tool of this scale. Until then, these roadblocks must be answered by either legal professionals or legislatures by updated laws in the near future.