We’ve already seen countless ways in which AI helps speed up medical diagnosis, and now it has a new trick: the ability to detect type 2 diabetes in someone, based on just a few seconds of conversation.
When having a conversation when you are emotional, drunk, or sleepy, you may notice that someone’s voice can be affected by a number of biological factors; This is a fact that opens the opportunity for artificial intelligence to detect subtle changes that can occur with changes in health.
It should be noted that the study was conducted by scientists at Klick Labs, who are interested in developing and selling AI detection technology. However, their findings have been published in a specialist journal and it is worth taking a look to see if type 2 diabetes detection can be improved.
“Current detection methods can be time-consuming, travel-intensive and expensive,” says Jaycee Kaufman, a research scientist at Klick Labs. “Voice technology has the potential to completely remove these barriers.”
The team asked 267 participants — some with type 2 diabetes and some without — to record a fixed phrase six times a day on a phone app over the course of two weeks. A total of 18,465 recordings were then processed to extract 14 different audio features, including pitch and intensity.
The researchers used a combination of these recordings to train the AI on a person’s voice, based on factors such as gender, age, body mass index, and whether or not they had type 2 diabetes. They used the remaining samples to test what the AI had “learned.”
Taking into account considerations such as age and gender, the model was able to detect type 2 diabetes with an accuracy level of 89% for women and 86% for men.
Interestingly, the main vocal signs that identify type 2 diabetes were different between men and women. In men, variation in intensity and amplitude was more significant; For women, differences in tone were the main revelation.
Researchers acknowledge that larger, more diverse groups of people need to be tested to validate these results, but early results are positive. Currently, the diagnosis of type 2 diabetes requires a blood draw, followed by a long wait for the analysis and report. This method requires little more than access to a smartphone app.
Although one in 11 adults worldwide are diagnosed with the disease, researchers believe that hundreds of millions of people do not realize they are living with type 2 diabetes. Being able to reduce this number also means being able to implement treatments earlier and reduce Costs. Managing diabetes in populations.
“Our research highlights significant vocal differences between individuals with and without type 2 diabetes, and could change how the medical community tracks diabetes,” Kaufman says.
Translated by Matthews Lineker from Science Alert
“Wannabe internet buff. Future teen idol. Hardcore zombie guru. Gamer. Avid creator. Entrepreneur. Bacon ninja.”