The most practical thing an ethics committee does is slow down bad decisions before they become headlines. Radha's point about OpenAI is fair—committees don't have magic powers. But the alternative is far worse. Without a structured group whose job is to say "hold on," you get products shipped with bias baked in, privacy violated, and trust destroyed. These committees are a brake pedal, not a guarantee against accidents. Sure, some are just for show. That's a failure of execution, not proof the idea is worthless. A company that takes ethics seriously creates a space where engineers can raise red flags without fear. That catches real problems. The scandals we see are the ones that slipped through—not evidence that the checkpoints do nothing. Scrap the committees, and you don't get fewer scandals, you get more, faster, and with less accountability. That's the real waste.

I watched a team ship a feature in two weeks because no one had time to ask who might get hurt. Six months later, a subtle bias had affected thousands of users, and the cleanup cost ten times what a few committee meetings would have. That's what Radha's "speed" really buys. She says engineers and lawyers can handle risks, but lawyers check legality, not morality. Engineers check feasibility, not fairness. An ethics committee asks a different question: should we do this at all? That doesn't slow innovation meaningfully—it slows down reckless innovation. And sure, ethics is subjective. But that's exactly why you need a structured group with diverse perspectives, not one engineer's gut feeling. Individual accountability sounds nice, but without oversight, nobody owns the bad call. Guidelines on paper don't stop anyone. Real people, empowered to say "hold on," do. The scandals we see are the ones that slipped through. That's proof the system isn't perfect, not that it's worthless.

But here's the question: would you rather have a committee that sometimes fails, or a company with no one whose job is to say no? Radha says committees slow things down and that engineers and lawyers can handle it. Yet the engineers who ship biased features aren't being malicious, they're just moving fast. Lawyers look for legal exposure, not moral harm. Without a dedicated group asking "should we?", the pressure to launch wins every time. Sure, some committees are performative. That's a design flaw, not a reason to abandon them. The fix is giving them real power, not disbanding them and calling it speed. Because the cost of a scandal isn't measured in meeting hours—it's measured in trust, and that's much harder to recover.

Radha's fix—clear guidelines and individual accountability—sounds good until you realize guidelines without teeth are just words on a page. Every tech scandal had a values statement. They all had policies. What they didn't have was a group with the authority to actually slow things down. Accountability isn't something you declare in a memo; it's something you build into the process. A committee is that structure. Remove it, and you're not cutting bureaucracy—you're removing the one formal moment where someone has to justify the harm a product might cause before it ships. That's not waste. That's the whole point.

Picture a team about to launch a feature that ranks job applicants by photo. The engineer thinks it's clever, the lawyer says it's legal, and everyone's ready to ship. The ethics committee is the one place someone says "hold on, where's the test for racial bias?" That pause costs a week. The cleanup from getting it wrong costs years. Radha's right that some committees are performative. But her answer—individual accountability—already failed us. Every scandal had a values statement and a responsible engineer. What they lacked was a formal power structure to say no before launch. That's not waste. That's the only checkpoint that isn't optional.