Free China Movement has published From the Great Firewall to the Great Model Firewall: AI, Censored Knowledge, and the Economics of Frontier Innovation in Mainland China, Working Paper 1.0 in its research series on information control, technological development, and frontier innovation.
The paper asks a narrower question than whether information controls prevent technological progress. Mainland China is already a major AI and innovation center. Stanford’s 2026 AI Index reports that leading PRC and U.S. systems have drawn close on several benchmark measures, while the PRC leads in a number of publication, citation, and patent indicators. The new paper therefore asks whether information restrictions can coexist with high technical capability while still narrowing particular parts of the knowledge-search space available to researchers and AI systems.
The paper distinguishes five stages at which such effects could arise: network access, training data, alignment and governance, retrieval and inference, and downstream knowledge production. Its principal empirical question is whether differences documented on explicitly political questions remain confined to those topics or extend into legitimate, policy-adjacent scientific and economic inquiry.
To test that proposition, Working Paper 1.0 sets out a 100-prompt, five-tier audit, together with same-weight deployment comparisons, source-diversity and omission measures, and controlled hypothesis-generation tasks. The paper also states in advance what would weaken its thesis: if politically sensitive differences do not extend into adjacent research domains after capability, language, and deployment are controlled, the broader innovation claim should be narrowed.
The study also examines Free China Movement’s proposed Third Republic transition architecture, but only as an institutional counterfactual. It does not assume that a new political order would necessarily produce better technological outcomes. Instead, it asks whether specific institutional changes—more reviewable information rules, wider research access, stronger legal predictability, and greater market contestability—could broaden legitimate knowledge search without sacrificing mainland China’s existing scientific, industrial, administrative, and human-capital strengths.
The new paper is a companion to The Innovation Tax of the Great Firewall: Information Control, Economic Catch-Up, and Mainland China’s Transition to a Frontier Economy.
Read Working Paper 1.0: From the Great Firewall to the Great Model Firewall.