'Randomizers' against profiling?

Just came across an article in LA Times entitled "Randomizers' could ward off airport profiling accusations" stating that one of the reasons for deploying them "is so TSA officers can't be accused of profiling passengers when they direct some fliers to a line for regular screening and others to a line for a faster, less-intrusive search". This is peculiar, I thought.

Indeed, the principle of using a random selection process as one of many layers in aviation security whereby a certain percentage of passengers is diverted to more intensive screening is not new (in fact it has been used by CATSA in Canada for years). It adds a certain degree of unpredictability in the system and makes it more difficult for individuals with malicious intent to circumvent the screening process. As the levels at which randomizers make their choices can be adjusted to lower or higher rates (there is an algorithm to do just that), the enhanced screening lines can process fewer or more people depending on threat level.

The problem with this approach could be...in its randomness. If a national security operator deploys random selection beyond a very small percentage (say 0.02 of purely random selection) it could be a sign of a system that chooses optics instead of a true risk-based approach to security where solutions are deployed based on threat/risk assessments. If a qualified TSA officer makes an informed decision to send a passenger to an enhanced screening line, the system behind that officer must stand firm and maintain that it is done based on a real-time risk assessment. That is, of course, if the system actually believes in the officer's knowledge and skills to perform the job right and operates based on a risk-based model.

Otherwise, unfortunately, it appears to be a smart communications ploy only.