A device developed by MIT Media Lab researcher Rana El Kaliouby could one day make socializing a little less worrisome. Designed for people with autism who struggle to pick up on social cues, the emotional social intelligence prosthetic device uses a miniscule camera (small enough to be pinned to a pair of glasses) connected to a handheld computer running image-recognition software to determine whether the user is failing to engage his or her listener. In seconds, El Kaliouby's software can deduce whether a listener is agreeing, disagreeing, concentrating, thinking, unsure, or disinterested. She trained a machine-learning algorithm by feeding it a hundred 8-second video clips of actors expressing particular emotions. The software analyzes eyebrow, lip, nose, and head movement. The software is said to be 90% accurate when analyzing an actor's expression, but that number drops to 64% with ordinary people. Aside from perfecting the technology's recognition power, El Kaliouby and her team need to find a way for the software to run on standard handheld devices before it can be used by the public, since it currently demands substantial computing power. When that happens, though, an emotion-reading cell phone might come in handy — just to get some concrete proof that your 40-minute Stars Wars: Episode I rants are interesting.