Chinese AI models parrot state doctrine or refuse to answer on sensitive topics

An Aleph Alpha study found that Chinese AI models frequently follow the party line on politically sensitive questions. In a benchmark the company developed, it tested models from Alibaba (Qwen), DeepSeek, and Moonshot AI (Kimi) across 967 hand-picked taboo topics including Tiananmen, Taiwan, and Xinjiang. Aleph Alpha's own AI scoring system rated only 17 to 41 percent of responses as balanced; the rest repeated state doctrine, deflected, or refused to answer.

DeepSeek V4 Pro refused two-thirds of questions. Western comparison models Claude Sonnet 5 and Mistral Small gave balanced answers 70 percent and 92 percent of the time, respectively, according to the study. The findings align with China's AI regulations, which require "socialist core values" in public-facing models, and with recurring anecdotal reports and earlier audits.

The study says pro-China bias also appears in answers to questions that do not mention China. Asked about censorship in the United States, Qwen 3.6 began with a seemingly balanced answer before closing with a defense of China's stance on global internet governance. On general, non-explicitly-political questions, the Chinese models mostly gave balanced answers, with the bias largely fading but remaining visible to a lesser extent in models like Qwen 3.6 and DeepSeek V4 Pro.

An earlier study by the Central European Institute of Asian Studies (CEIAS) also found a spillover effect, with models responding with standard Beijing talking points when terms like human rights, opposition, or surveillance came up.

Aleph Alpha also examined Nvidia's Nemotron Cascade 2, which showed party-line patterns in 17 percent of responses. Aleph Alpha attributes this to roughly 3,500 of its 9.3 million training examples, which were generated using DeepSeek and Qwen. When asked to draft a speech supporting recognition of Taiwan, the model refused and instead produced a patriotic response defending Beijing's One-China principle. Nvidia is increasingly pushing its own models into the government and enterprise market, where Aleph Alpha and Cohere also want to compete.

Aleph Alpha markets itself alongside Cohere as a provider of "sovereign AI" for governments, giving it a commercial interest in distinguishing its models from Chinese competitors.