Lie Detectors used in call centres

There are a few stories of stress detection (read lie detection) software being used in call centres to help combat fraud. Currently, the software has been trialled in the UK to detect benefit fraud cases and is now being rolled out into production. The software uses a type of technology called voice risk analysis (VRA) that detects subtle changes in the voice to quantify the risk that the caller might be lying.
The software compares truthful answers to answers given to questions where the caller might lie. If differences are detected, the caller is identified as a risk (i.e. considered to be lying), and some other process is used to verify the facts of the caller. There have been mentions in the past of insurance companies using the technology for claims too. One provider of the technology that keeps popping up is DigiLog who explain the process in more detail on their website.
I don’t have an issue with the software being used to guide the process of combating fraud. However, I know from experience that most operators in call centres are lazy and are most likely to rely on the designated risk for the caller as being the absolute decision if the caller is lying or not. There’s a huge amount of discussion about the subject and many experts consider the technology to be inconclusive (read unreliable, rubbish and error prone).

(2 votes, average: 4.5 out of 5)




