IBM and Amazon recently announced they are stepping back from Facial Recognition programs.
IBM is stopping research and development, Amazon is taking "a moratorium" from using or selling to police forces - except for cases of human trafficking.
Very interesting and this is consistent with concerns that were being raised about facial recognition technology going way back.
Facial recognition was actually a module on my course at university. You may be surprised to know that it goes back to the 1960s, by a system pioneered by a chap named Woodrow Wilson Bledsoe whose device used horizontal and vertical co-ordinates on a grid using a stylus that emitted electromagnetic pulses. The system could be used to manually record the coordinate locations of various facial features including the eyes, nose, hairline and mouth.
During the 1970s the accuracy of this system was increased to be able to specify 21 facial markers. In the 80s and 90s the "Eigenface" technique showed that using feature analysis alone it was possible to create a facial image from scratch..!!
In the 2000s, the FERET Programme created a database of more than 2000 images that would be used for test purposes as the technology progressed. At the 2002 Superbowl, fans were unknowingly scanned using this database in the hope it would "Prove the system." It failed. I remember attending a lecture on this very experiment in 2004. The Prof who gave the lecture was quite scathing about it.
Of course, none of these things has stopped research and development of FR technology. And if it is used for benign purposes it can be very useful. However, it is mostly law enforcement organisations that are most interested, eg, at airports and in detecting criminals. Fair enough as long as it is used properly and doesn't infringe human rights or civil liberties.
And that's where it starts to go belly-up. Such systems are wide open to malpractice and abuse.
And that, my friend, is the whole point of this discussion.