The software should be able to speed up the time in takes to separate truth from fiction.

Politicians have traditionally been very good at masking their emotions when it suits them, but we humans are easy to fool. A machine, on the other hand, might be able to peer through the shroud of deception to get at the true nature of a politician's emotional state. Or at least that's the goal of the project undertaken by four grad students at Columbia University. They built an AI application called "Debate in (E)motion" to watch the recent US Presidential debates to evaluate the emotions of each candidate. The AI isn't listening to what the candidates are saying, and it has no political bias. It is simply extracting a frame from the video every five seconds and scanning for recognizable human emotions.