Detecting lies with text mining software in not only a possibility but a reality. However, the reliability of such a method still remain to be improved. Work has been done on this topic since more than 30 years and more recently, James W. Pennebaker from the University of Texas use his content analysis software (LIWC) to detect deception. If you are interested in reading on this topic, you can look at the last two references mentioned on this page:
Credit card fraud detection has been done by computers for a number of years.
Recent articles about text mining software raised the possibility that software can read emails or other written text and detect lies.
Do you think it is possible to do this in the next 5 years?