Raw LLM Responses

It's not news because people are bad at statistics, you just don't understand why people are upset. You are correct that this is an *extremely* accurate system from a statistical and technological perspective. 99.5% accuracy is quite good. But you're still wrong. The fact remains - the overwhelming majority of people flagged were false positives. This *isn't* an argument that the system is flawed - it's doing what it's designed to do and doing that pretty effectively. It's an argument against sweeping facial recognition based mass surveillance entirely. You're mistaking a moral argument for a technological/statistical one. In fact, what you're saying drives the point home even more: the system is working quite well and doing what it's supposed to do and what the police wanted it to do. Yet in spite of that, 81% of results are false positives. Those are real human beings with rights too. It's a little depressing that you've posted a convincing argument for why any sort of large scale automated mass surveillance is inherently repugnant, only to completely miss the point.