That is unironically more sovereign than using any US-developed AI. GLM 5.3 is cutting-edge for cybersecurity tasks, while the US frontier forces KYC/background checks for cyber capabilities.