FlyerTalk Forums - View Single Post - Can Robots Better Spot Terrorists at Airports?
Old Jan 1, 2014, 12:11 pm
  #6  
jkhuggins
 
Join Date: Nov 2008
Posts: 3,657
West, you know the respect I have for you --- especially for braving this forum on a regular basis. But I'm going to disagree with you pretty strongly on this one.

Originally Posted by gsoltso
Some of this includes watching people for those that stand out - you and about 99% of us do this all day, everyday, everywhere. You get that sense of someone that you don't want to be near, you avoid them, many times subconsciously.
And, many times, that subconscious judgment is based on items that are not appropriate for a security officer --- judgments based on race, on gender, on religion, and so on.

Originally Posted by gsoltso
Police Officers do it all day while on patrol, something sticks out, they (at least some of them do) go check out what is going on.
And look at how much controversy "stop-and-frisk" has in the communities where it is practiced, such as New York City. Most NYPD officers would tell you that they're simply checking out people who "don't fit" ... and, lo and behold, the percentage of African Americans who "don't fit" far exceeds their proportions in the general population.

Originally Posted by gsoltso
I think that one problem with automating the system completely is losing that human element of intuition.
It also potentially eliminates that human element of prejudice.

Originally Posted by gsoltso
We can program the crap out of computers, but until we have some pretty serious software breakthroughs, they are not going to have the same natural reaction to things that make us aware of someone that is outside of the norms and mores of the location.
Ah ... the thoroughly-discredited "War on the Unexpected".

Originally Posted by gsoltso
While I am all for automation as part of the process, I would always be more comfortable with some humans at the location (regardless of whether it remains TSA or some other form of security as many suggest). We have developed this internal "radar" over millenia, to ignore it is to deny an effective tool in the process.
As a technologist, I'm in agreement that security screening tasks shouldn't be completely automated. Technologies fail, and humans are needed to recognize their failures and compensate for them. But let's not pretend that humans are perfect either. There are times that impersonal judgments are to be preferred over subjective human judgments.
jkhuggins is offline