We should have learned from navigation systems that people rely too much on a computer without thinking for themselves.
Facial recognition and AI for behavior recognition are potentially good security tools. But only if a human controls the output. The US had people wrongfully imprisoned because some Flock camera made the wrong call. Cases in which a human would have noticed that the accused’s car has an entirely different color.
It’s a human problem that can’t get solved. Not with new protocols, not with new software ideas. I have seen people clicking on buttons without reading what is there, only to get rid of any obstacle fast enough.
The only solution is banning this technology. Alternatively, make all false positives not caught by a human as expensive as possible.
Richard, Elliot’s Assistant: This is pure Kafka…
Hospital Interrogator: Who’s Kafka! Tell me!
https://www.imdb.com/title/tt0112715/quotes/?item=qt0429507