Super Bowl XLIII will be the first time that federal behavior-detection agents will be used at a major event, reports USA Today.
At the Tampa Police request, the TSA is sending dozens of its behavior officers to Tampa to watch spectators entering 75,000-seat Raymond James Stadium on Sunday, said Tampa Police spokeswoman Laura McElroy. […]
Behavior observation is used daily by 2,600 specially trained TSA officers at more than 160 airports. Officers look for obvious signs of nervousness or other behavioral flags, such as sweating, avoiding eye contact or talking evasively.
There are a lot of privacy and civil liberty questions surrounding behavior detection programs.Â There are any number of innocent reasons why an individual would be nervous or agitated at an airport or the Super Bowl. Wouldn’t you presume nervousness from someone who spent that much money and time to go to the Super Bowl to watch his or her team play?
What would the error rate of this technology be? How many false positives leading to the harassment of innocent people and the diversion of investigatorsâ€™ attention and resources from actual criminals?
Questions about behavior detection programs aren’t just coming from privacy advocates. In October, the National Research Council released a report,Â â€œProtecting Individual Privacy in the Struggle Against Terrorists: A Framework for Program Assessment,â€Â which was sponsored by the Department of Homeland Security and the National Science Foundation. The authors heldÂ a panel discussionÂ on the reportâ€™s findings.Â
One important point from the panel discussion: Any programs attempting to assess an individualâ€™s state of mind is considered suspect. Such behavior detection programs should be a tool for further investigation, but not determinative of intent because there is a high probability of false positives.
Behavior detection programs are proliferating, and detection technology is being tested in the US and abroad. In September,Â USA TodayÂ discussedÂ new behavior detection technology being developed by the Department of Homeland Security; it seeks to divine an individualâ€™s criminal or benign intent from a bio scan. Also that month,Â New York TimesÂ reported on a caseÂ in India concerning brain scan technology believed to identify when an individual is lying. The brain scan evidence was used to convict a woman accused of murdering her fiance.
I have blogged aboutÂ the European Union testing in-flight video surveillance to detect criminal intent. Previously, I wrote anÂ op-edÂ (pdf) inÂ The TennesseanÂ about face recognition systems being used in schools. I explained that the algorithms to automatically detect and identify individuals were complex and still error-prone.
A 2007 large-scale face recognitionÂ scientific studyÂ (pdf) found that, during the day, the error rate of the technology was about 40 percent. At night, the error rate wasÂ 80 to 90 percent.