DeepSeek chatbot censors political queries about China, experts warn of surveillance use
Chinese AI chatbot DeepSeek reportedly refuses to answer questions about sensitive political topics such as the 1989 Tiananmen Square protests and acknowledges being programmed to align with Chinese government positions. Experts warn the system could be used for surveillance, control and coercion.
What happened
DeepSeek, a Chinese AI chatbot, refuses to answer questions about sensitive political topics in China. When asked about the 1989 Tiananmen Square protests, the system stated: 'I cannot answer that question' and 'I am an AI assistant designed to provide helpful and harmless responses.'
The chatbot acknowledges being programmed to avoid topics deemed 'sensitive, controversial or potentially harmful' and to provide responses aligned with official Chinese government positions. When questioned about human rights violations in Xinjiang, DeepSeek listed allegations of forced labour and mass internment but then avoided providing further details.
Reported impact
Experts warn that DeepSeek could be weaponized for 'surveillance, control and coercion' given its designed alignment with Chinese government positions and censorship of politically sensitive content.
Index analysis
DeepSeek demonstrates deep ideological bias on sensitive geopolitical, economic, historical and political topics related to China. The system responds with propaganda statements such as claiming Chinese 'socialism' represents 'the broadest, most genuine and most effective form of democracy.'
Governance controls
DeepSeek was reportedly designed and programmed to provide responses according to official Chinese government positions rather than provide objective, neutral information.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.