Dario Amodei, CEO of Anthropic, has proposed that governments intervene to slow the development of powerful AI models. His plan includes embedding independent evaluators within AI companies and establishing international agreements with China regarding testing and safety.
China has already implemented a multi-stage regulatory framework for artificial intelligence. In 2022, Beijing introduced rules for algorithmic recommendation services, followed by 2023 provisions for deep synthesis and interim measures for generative AI services.
The latest regulatory step is the AI safety governance framework 3.0, published in September 2026. It establishes mechanisms for risk identification and management across the AI lifecycle, including autonomous systems and AI agents.
These Chinese regulations mandate that companies perform security assessments, file algorithm information with authorities, and ensure the quality of training data. The framework also addresses risks related to chemical, biological, and nuclear technologies.
While US leaders debate the need for future oversight, China has already integrated evaluation, technical standards, and government supervision into its AI sector. The US currently maintains a market-led approach with minimal regulatory burdens.
