Computers can Catch Fake Pain Better than HumansCategory: Science & Technology
Posted: March 24, 2014 01:32PM
In today's culture, much of our communication utilizes written or more accurately, typed words in the form of tweets, updates, posts, and more. Throughout and even before human history though, our culture evolved with communication that used more than words, but also physical cues. Researchers at the University of Toronto have recently built a computer system capable of more accurately indicating when facial cues are false than actual humans.
With humanity's ability to read a person's emotions from facial cues, we have also developed ways to trick each other by putting on faces, not representative of our situation. When tested it was found that the average person is no more accurate about if someone is in pain or faking, than if they guessed. Even with training the numbers do not get much better, but the computer system was accurate 85% of the time. One of the primary indications of if someone was faking was how the mouth behaved, and if it opened and closed too regularly.
While this study may seem to do little more than indicate a sci-fi future controlled by robots, it could have some important, real impacts. These include systems to catch other deceptive actions than faking pain and perhaps could develop into ways to track how sleepy a driver is, how responsive treatments are for some disorders, and if students are actually paying attention during lectures.