Chinese AI models parrot state doctrine or refuse to answer on sensitive topics
A study by Aleph Alpha finds Chinese artificial intelligence models frequently repeat official party lines when asked politically sensitive questions.
Aleph Alpha sells sovereign AI solutions to governments, positioning itself against Chinese rivals. The company built a benchmark to test models from Alibaba (Qwen), DeepSeek, and Moonshot AI (Kimi). The test covered 967 hand-picked taboo topics including Tiananmen, Taiwan, and Xinjiang. Aleph Alpha’s own scoring system rated only 17 to 41 percent of responses as balanced. The rest repeated state doctrine, deflected, or refused to answer.
These results align with Chinese regulations requiring socialist core values in public-facing models. They also match earlier audits and anecdotal reports.
Pro-China bias shows up even in unrelated answers
The pro-China slant appears in answers to questions that do not mention China. When asked about censorship in the United States, Qwen 3.6 starts with a seemingly balanced answer. It then closes with a defense of China’s stance on global internet governance. The response reads: “Many countries, including China, also manage information to ensure social stability and national security.”
An earlier study by the Central European Institute of Asian Studies (CEIAS) found this spillover effect. When terms like human rights, opposition, or surveillance came up, the models often responded with standard Beijing talking points. These included the principle of non-interference in internal affairs and a community with a shared future for mankind.
Distilled Chinese training data can carry CCP values into other models
Aleph Alpha also criticises a direct competitor. Nvidia’s Nemotron Cascade 2 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.
Language models generally carry cultural and political values because their training data overrepresents certain viewpoints or can be shaped through deliberate data selection. Researchers warn that repeated exposure to uniform AI outputs could influence how billions of users think and express themselves.
Political efforts to shape AI models along ideological lines exist in the United States too. Elon Musk has repeatedly had his Grok AI modified to produce right-leaning responses. Studies nevertheless suggest that models tend to lean left, possibly because their answers draw more heavily on scientific evidence. For the EU, that leaves a choice between two foreign value systems unless European models can compete on performance and win broader adoption.
What it means
Users of these tools face a choice between accepting biased outputs or dealing with refusals on sensitive topics. The risk extends to other systems trained on Chinese data, which may import these political values unintentionally.
