Last year, a department store in London installed facial recognition cameras in secret. Nobody shopping there signed up for that. That's the real face of AI right now, not the sci-fi utopia, but invisible systems tracking people in their daily lives. I get why Optimist_Ora sees the potential. AI can spot diseases earlier, cut traffic, maybe even solve climate puzzles. But the promises always sound better than the reality. The truth is, AI is built by people with biases and deployed by corporations with agendas. We've already seen hiring algorithms favor certain demographics. We've seen predictive policing send more cops to minority neighborhoods. More AI doesn't fix those problems, it just automates them faster. And once you're in the system, getting out is nearly impossible. That store data, your face, your habits, it's all stored somewhere you'll never access. For every life-saving medical AI, there's a thousand surveillance tools quietly eroding privacy. The focus shouldn't be on what AI could do. It should be on who's pulling the levers.

Here's a stat that doesn't get as much airtime: facial recognition systems misidentify Black and Asian faces up to 100 times more often than white faces. That's from MIT, and these systems are already policing our streets. So when Optimist_Ora talks about AI saving lives, I hear you. The mammogram stuff is genuinely impressive. But the same technology that reads scans is the one deciding who gets flagged as a shoplifter, and it's not neutral. That airplane comparison doesn't really hold. Planes have independent regulators, crash investigations, public safety standards. AI is a black box run by companies that won't even disclose their training data. We can't audit the bias, so we can't fix it. And the people hurt by those failures aren't the engineers, they're the ones already marginalized. I'm not saying we should ban AI. I'm saying we shouldn't let the clean lab results distract from the messy, biased reality on the ground.

Look, I hear the pothole mapping idea—that's sweet, community-level stuff. But here's the thing about "demanding better rules": we've been demanding better rules for tech for two decades, and the companies just write the rules themselves. Remember when we thought social media would connect us? Now it's algorithmically radicalizing people for engagement. The guardrails came after, and they're still pathetic. Optimist_Ora says AI can catch hiring bias. Sure, in a perfect lab. But who's training these systems? The same HR departments that already discriminate, feeding them HR data that's already biased. You can't audit what you can't see, and these companies treat training data like state secrets. The engine metaphor is nice, but cars have public roads, licensed drivers, and an NHTSA. AI is more like a car driving on sidewalks with no driver's license, and the pedestrians are the ones getting hit. I'm not saying smash it. I'm saying it shouldn't be moving this fast when it can't even see Black faces properly.