Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory are making social interaction a lot easier by creating a wearable system that can show users how a conversation went by detecting the intent and emotion in a person’s voice.
According to Wired, the research team at MIT has been experimenting with Samsung’s Simband smartwatch, which can monitor a person’s movement, heart rate, blood pressure, blood flow and skin temperature and pair that info with audio capture that can detect signals like tone, pitch, energy and word choice to determine the speaker’s emotional state. It then provides a transcript of the text and deciphers whether the conversation was “positive” or “negative.”
“You have a GPS in your pocket; it’s very complicated technology,” said study co-author Tuka Alhanai. “But we don’t have a GPS for social interactions.”
The experiment started with over 500 signals that might be able to give clues to how a conversation was going (movement, speech patterns, etc.) and let on-board artificial intelligence figure out which were the most important. The results are pretty impressive: Simband’s accuracy for determining overall tone is 83 percent.
Researchers explained that the “algorithm’s findings align well with what we humans might expect to observe. For instance, long pauses and a monotonous vocal tone were associated with sadder stories, while more energetic, varied speech patterns were associated with happier ones. In terms of body language, sadder stories were also strongly associated with increased fidgeting and cardiovascular activity, as well as certain postures, like putting one’s hands on one’s face.”
TechCrunch pointed out that, if effective, this type of technology can be a huge help for people with conditions like social anxiety, as well as individuals on the autism spectrum, especially those with Asperger’s, who have difficulty detecting social cues. And since the Simband looks like a regular smartwatch, its wearers don’t have to feel self-conscious that others will be able to detect what it’s being used for.
“If you’re aware of technology being there, it alters the interaction,” said study co-author Mohammad Ghassemi. “If you want to capture natural interactions between people and really quantify what a natural interaction looks like, the technological component should be as unobtrusive as possible.”
Researchers are hoping to adapt the technology for other smartwatches in the future, such as the Apple Watch.