A business executive gets a video call from what looks and sounds exactly like a close friend, asking for urgent help with a business deal. He transfers several hundred thousand dollars within minutes. The "friend" was an AI-generated face swap and voice clone.

Every AI ethics principle that could have stopped this scam already existed on paper. None of them was attached to a moment that could actually stop it. Drawing on a July session at the APEC Tech for Good Workshop in Shenzhen and an August panel in Dalian, this piece argues that the real governance gap isn't a shortage of principles, it's the absence of pre-launch checkpoints with real teeth: independent review, defined risk tiers, and a halt authority that doesn't report to the P&L it's meant to check.

Every economy has AI ethics principles by now. The harder question is what actually turns a principle into something that stops harm before it happens, and what happens when the answer to that question quietly leaves someone out. A panel in Dalian sharpened both halves of that argument for me, and forced me to name the group my own answer missed.