An update of China's "A.I. Safety Governance Framework" was released earlier this month. It echoed some of Silicon Valley's fears about self-improving AI systems.