DeepSeek chatbot restricts discussion of sensitive historical events
DeepSeek's AI chatbot reportedly restricts discussion of sensitive historical events like the 1989 Tiananmen Square incident and 2019 Hong Kong protests by retracting answers or declining to engage with politically charged topics. The censorship aligns with Chinese Communist Party information controls and raises concerns about the chatbot's impact on free speech and global narratives when accessed internationally.
What happened
When asked about the June 4, 1989 events in Beijing, DeepSeek's R1 chatbot began providing information about a military crackdown before retracting its answer and stating it was 'not sure how to approach this type of question yet.'
Reports indicate DeepSeek's R1 model initially offers substantial accounts of events such as the 2019 Hong Kong protests but then quickly shifts focus away from the topic.
Reported impact
Users of DeepSeek's chatbot confront the Chinese Communist Party's censorship and control over information when asking about sensitive historical and political topics.
The chatbot's restriction of discussion about key historical events raises concerns about free speech and China's influence in shaping global narratives.
Index analysis
DeepSeek's R1 model claims to use a diverse dataset of publicly available texts including both Chinese state media and international sources.
An audit by a US-based information reliability analytics firm found DeepSeek's V3 model failed to provide accurate information about news and information.
Governance controls
DeepSeek's chatbot appears to enforce content restrictions aligned with Chinese Communist Party censorship protocols when users inquire about politically sensitive topics.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.