Lie-detecting algorithm spots fibbing faces better than humans

THINK twice before you tell that fib. By watching courtroom videos, a computer has learned to predict if someone is telling the truth or a lie. A machine learning algorithm trained on the faces of defendants in recordings of real trials, including that of Andrea Sneiderman (below) who was convicted of lying, correctly identified truth-tellers about 75 per cent of the time. Humans managed just 59.5 per cent. The best interrogators can reach 65 per cent. “We’re actually pretty bad lie detectors,” says Rada Mihalcea at the University of Michigan in Ann Arbor. Mihalcea and her colleagues took 121 videos from sources such as the Innocence Project, a non-profit group in Texas dedicated to exonerating people with wrongful convictions. This is superior to simulated conversation because the speakers are more…


Link to Full Article: Lie-detecting algorithm spots fibbing faces better than humans