Researcher developing first online polygraph test
Marie Donlon | March 21, 2019A Florida State University researcher is developing an online polygraph test that identifies when someone is lying based on cues in their electronic messages.
Shuyuan Ho, an associate professor in the College of Communication and Information, has created an online game that gauges both truthful and deceptive communication between two players. By parsing the words exchanged in those conversations, Ho extracted context behind millions of bits of data in the messages that served as so-called “language action cues,” which are similar to the physical cues a person might betray when telling the truth or lying. During the game-playing experiment, Ho discovered that participants could detect lies in messages roughly 50% of the time while the machine learning approach was able to identify deception with an 85% to 100% rate of accuracy.
Participants playing the game were randomly assigned to either the role of “The Sinner” or “The Saint.” While playing the game, researchers captured the conversations held between The Sinner and The Saint and applied the machine learning technique to the communications to examine word and writing patterns. Apart from word choice differences, researchers discovered that The Sinners, or liars, tended to respond to messages in less time than their truth-telling counterparts The Saints who took longer to respond to queries.
Noticing the distinction, researchers time-stamped the lapses within and between sentences to measure the duration of pauses. Considered a language action cue, some of the pauses were so short that they could escape notice were it not for machine learning capturing the lapses.
Once fully developed, Ho envisions that the online polygraph test could be put to use against online deception such as trolling, identity theft and phishing schemes.
"You could use it for online dating, Facebook, Twitter — the applications are endless. I think the future is unlimited for an online polygraph system," said Ho.
"I want to get the world's attention on this research so we can hopefully make it into a commercial product that could be attached to all kinds of online social forums," Ho said. "This basic research offers great potential to develop an online polygraph system that helps protect our online communication."
The research is detailed in the journal Computers in Human Behavior.
Too easy to train yourself to avoid detection once you understand how the algorithm works.
Trying to modify innate physiological responses is more difficult. Even a full up polygraph depends in large part on the skills of examiner in concert with the telemetry from the measuring equipment. The process of designing baseline questions is fraught with ambiguity because typically those baseline questions are rather innocuous, so the responses will be different than a 'charged' question in which the question itself generates a physiological response in the examinee.