Great new security tech, or fraud?
While at CanSecWest, I was noting a news story about how somebody had, yet again, defrauded the US government and military by selling them a terribly sophisticated computer algorithm that promised to find secret information about enemies and/or terrorists, but actually didn’t work. I suspect that this will be a complex case, since the vendor will undoubtedly claim that his work is so sophisticated and complicated that it does work, it’s just that the users didn’t understand it.
In view of this, I found it really interesting to note a very similar case, just a few days later. Computerized Voice Stress Analyzers (CVSAs) have been promoted and sold for a least 25 years now. This despite the fact that, four years ago, the U.S. Department of Justice did a study and concluded that “VSA programs show poor validity -neither program efficiently determined who was being deceptive about recent drug use. The programs were not able to detect deception at a rate any better than chance … The data also suggest poor reliability for both VSA products when we compared expert and novice interpretations of the output.”
In a sense the CVSA case is much worse, because, since it is a private company selling to private companies, there is nobody to say that these people are a) wasting money, and b) making poor hiring decisions based on what is essentially a coin flip.