China and the US should not only enhance the capacity for risk assessment and safety supervision in AI, but also help developing countries in the field