Raw LLM Responses

Before I offer my opinion as a computer vision expert, let it be known that I find the idea of universal surveillance abhorrent and am a staunch supporter of individual rights including the right to mind my own fucking business without Big Brother looking over my shoulder. As someone who worked with the designers of a face recognition system for criminal identification (though MUCH narrower in scope, think “here’s a picture of the guy who robbed me, find his mugshot”), the only purpose should be to narrow down suspects. What we really want to know is how many people it flags per day, or the hit rate. Let’s say your house is robbed and you get a few good photos of the burglar on your security camera. The system can take those shots and look for people who match them and relay that information back to a human. The human then sifts through the data and narrows it down MUCH further. Instead of looking at millions of people per day, say the system flags 50. You only have to check to see if one of those 50 match the person you’re interested in. It’s also important that the operators be trained that a positive match doesn’t mean that’s your suspect, but rather that it *might* be them. Now, should this be deployed in a mass surveillance system? Hell the fuck no! It’s a breach of my right to privacy. I just wanted to clarify some of the terms used.