New Publication: Teaching Computers to Lie


Rowe and Rrushi’s
Introduction to Cyberdeception

Published in Enculturation,
December 15, 2017

Excerpt: “Suppose you are a criminal investigator interrogating a suspect. If you have reason to believe that your suspect is attempting to deceive you, then you will be on the lookout for certain telltale signs of a liar: vagueness, contradiction, hyperbole. Face-to-face, these marks of deception appear like a stain on a new shirt or like broccoli caught in the teeth. To your trained eye, the clues are impossible to ignore: sweaty palms, nervous fidgeting, lack of eye contact (it’s all too easy now, Detective Briscoe). If your suspect had watched more Law & Order, if she had learned to avoid the obvious missteps, she would perhaps have some chance at getting away with her deceptions.

Now suppose you are an investigator of cybercrime, looking for clues across digital space. Online avatars don’t fidget. Operating systems don’t have eyes (at least, not yet). The limited number of clues in cyberspace make it that much easier for criminals to deceive you. Yet according to Neil C. Rowe and Julian Rrushi, in their book, Introduction to Cyberdeception (2016), many of the traditional methods used to detect face-to-face lies remain applicable in online worlds. Such methods can be used to detect online scammers, suspicious algorithms, and shady software. Consider the Nigerian spammer who sends bulk emails replete with superfluous information, unexpected complications, incoherent logics. This sort of amateurish deception is generally easy to detect.” [READ MORE]