MIT Student Wants to Teach You How To Be Less Awkward in Public

A program developed out of the school helps "coach" people through their social phobias.

Sitting through job interviews, finding things to talk about on first dates, and mingling with random folks at parties and large gatherings can all be scary situations for a lot of people (social phobias affect roughly 15 million adults, according to the National Institute of Mental Health). But a new program developed at MIT is giving people some practice tips, to fine tune their social skills, before they go out in public, so they feel less awkward when interacting with others.

Called the “My Automated Conversation Coach,” or MACH, a 3-D computer-automated face that reads gestures of users when sitting in front of a screen, people can learn to become comfortable talking with strangers by improving skills in the privacy of their own homes.

According to MIT News, the program, designed at the MIT Media Lab, works on a personal computer and makes its own decisions based on the decisions of the user. By utilizing a web cam, the program can monitor facial expression and gestures, like when a person moves their head, smiles, or looks away, allowing for feedback based on their responses to the computer-generated image on the screen.

Users can even play back the video and calculate precisely where they smiled, how their voice changed during certain parts of the conversation, and when their attention may have wandered off during the course of the interaction. “In a study with 90 MIT undergraduates, the subjects went through simulated job interviews before and after receiving this training. Those who got feedback using this automated system were rated as better candidates for the job than those who did not,” according to a video posted by MIT News that details how the program works. “Besides job interviews, researchers say the system could help with public speaking, dating, learning languages, or helping people who have difficulties in social communications.”

The research was led by MIT Media Lab doctoral student M. Ehsan Hoque, who told MIT News that the reason it’s effective is because people won’t feel embarrassed getting feedback by a computer, as they would, say, if a person critiqued their social skills face to face. “It’s easier to tell the brutal truth through the [software] because it’s objective,” he said.

According to Hoque:

The system enables social skills training beyond counseling offices and clinical facilities, when and where users want it. The system affords embodied interaction and real-time feedback through a humanlike virtual character and postinteraction feedback on user performance through a visualization interface.

Hoque worked on the program over the span of two years as part of his doctoral thesis with the help of a group of advisors and undergraduate students. His work will be showcased at the 2013 International Joint Conference on Pervasive and Ubiquitous Computing.

Hoque said he hopes the program can help more than just people who lack social skills. “We wish to explore other application areas of our system for social skills training such as helping individuals with communicative challenges.”