Concerns over frontier AI models, such as Anthropic's Mythos highlighting cyberattack potential, contributed to Beijing's push for increased national control over AI development