Carnegie Mellon AI project brings social skills to virtual bots.
Artificial intelligence brings levels of intellect and reasoning not possible with the human mind. But what AI-powered virtual assistants such as Siri typically lack is the human ability to "read people," which ultimately limits their potential.
Carnegie Mellon’s Socially Aware Robot Assistant (S.A.R.A.) project aims to teach virtual assistants how to pick up on social cues—something the researchers claim has never been done before. The platform uses cameras and a mic to listen to people’s queries and take in how they are feeling as they interact, according to an article in Popular Science.The idea is that the assistant should respond differently to someone who is smiling and giving a fairly traditional answer versus another who is frowning and acing in a way that doesn’t comply with social norms.
The researchers are touting three new avenues of exploration: Conversational strategy classifiers, a rapport estimator, and a social reasoner. They are even exploring using S.A.R.A. for peer tutoring so if one student is struggling more than another, it will offer to lend a hand with extra social support.
Now that’s AI technology worth friending.