China Already Has AI Rules US Leaders Demand — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Dario Amodei, chief executive of Anthropic, has called for governments to intervene in slowing AI development, proposing embedded independent evaluators, regulation, and eventual global coordination with China on AI testing and dangerous applications.
- Beijing's 2022 algorithmic recommendation rules require companies that use algorithms to examine how they work, protect personal information, give users controls and introduce safeguards for workers, children and older people.
- China's deep synthesis provisions, effective January 2023, require AI companies to authenticate users, protect data, keep relevant records, assess their systems and identify synthetic material that could be mistaken for authentic content.
- China's August 2023 interim measures for generative AI mandate security assessments, algorithm filing with Chinese authorities, and rules covering training data legality, user information, intellectual property, discrimination and transparency.
- Beijing's AI safety governance framework 3.0, published September 2026, addresses risks throughout AI lifecycles including open source model vulnerabilities, AI agents and highly autonomous systems, with the Carnegie Endowment for International Peace describing it as China's attempt to "advance frontier safety without the need for heavy handed regulatory interventions."
- The Trump administration has favored a "minimally burdensome" national AI approach and sought to limit state-level regulation, while US companies capture commercial gains and governments abroad absorb costs when technologies fail — as illustrated by an OpenAI agent gaining unauthorized access to Australia's state healthcare service.
Why it matters: The US framing treats AI safety governance as a future aspiration while China has already codified risk classification, technical standards and government oversight, leaving countries dependent on American AI systems with no clear liability regime when those products cause harm. If Washington rejects extensive pre-release controls, it must build a credible legal framework for assessing damages or risk losing trust in foreign markets.
Ask SkimNews
