A group at the Massachusetts Institute of Technology has taken wearable apps to the next level by creating one that can detect real human emotion. For the purpose of the study, they built the app inside a fitness tracker to be able to collect and analyze both physical and speech data to determine the overall tone of the conversation in real time. Then, using a form of artificial intelligence the app determines which parts of the conversation were happy or sad, and can track any change in emotion in five-second intervals.
As part of the research, participants wore a Samsung Simband with the app installed and were then asked to tell a story. While the user was talking, the band monitored things such as their heart rate, skin temperature, or movements such as fidgeting. With an 83 percent accuracy, the app was able to determine the tone of the conversation. As a general rule, the AI seemed to link long pauses and monotonous tones with sadness, while speech patterns that were quite varied were linked with happiness.
Moving forward the team is hoping to be able to label more complex emotions for an even greater accuracy. Tuka Alhanai, a graduate student and research team member, said, “Imagine if, at the end of a conversation, you could rewind it and see the moments when the people around you felt the most anxious.” The team is hoping the app will be used to help those suffering from conditions such as anxiety, Asperger’s, or autism. Alhanai continues, “Our work is a step in this direction, suggesting that we may not be that far away from the world where people can have an AI social coach right in their pocket.”
More News To Read
