A U.S. startup, Canary Speech, is developing an algorithm that will be able to identify the symptoms of certain neurological diseases by analyzing the patient’s voice. The company has access to a huge trove of audio data comprising hundreds of millions of phone calls received by a major American healthcare and insurance provider over the past 15 years.
Canary Speech seeks to use this information to pick up cues that distinguish someone with a particular condition from someone without that condition, notes New Scientist.
The startup’s approach to use big data for recognizing subtle differences in the voices of people with different condition however remains open to question. Max Little at Aston University in Birmingham, U.K., says accumulation of data is not going to provide fruitful results.
The idea of vetting phone calls and then devising an algorithm than can be used to “screen” callers or influence insurance premium has also alarmed the observers but Canary Speech claims such applications “may be regulated.”