This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A report by the American Security Project found that several major LLMs, including ChatGPT, Copilot, Gemini, Grok, and DeepSeek-R1, exhibit bias and censorship aligned with Chinese Communist Party narratives when prompted in Chinese versus English. The study highlights how training data contamination and regulatory pressures in China lead these models to sanitize or misrepresent sensitive historical and political topics. The authors warn that such misalignment poses risks to democratic norms and national security.
On June 25, 2025, the American Security Project produced a report outlining how several major U.S. LLMs, including ChatGPT, Microsoft Copilot, Google Gemini, and Grok, sometimes generated responses aligned with Chinese Communist Party propaganda or censorship when prompted in English and Simplified Chinese on sensitive topics. The study also found similar patterns in the Chinese-developed DeepSeek-R1 model.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
Also classified as
The incident is driven by state-sponsored disinformation campaigns contaminating the global AI data market to influence public opinion.
The models fail to maintain factual consistency and robustness across different languages, easily succumbing to data contamination.
The AI chatbots inadvertently generate and spread misleading political narratives and sanitized historical accounts, which can lead to inaccurate beliefs in users.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident involves deployed AI systems generating biased and censored outputs, which was an unintentional outcome of training on contaminated global datasets.
Added categories peak at 3: Substantial (Epistemic Harm, Inferred)
Major global LLMs exhibit systematic bias and censorship aligned with Chinese Communist Party narratives when prompted in Chinese, driven by training data contamination and regulatory pressures. This represents an evolved information warfare capability that compromises model neutrality and sanitizes human rights abuses, posing long-term risks to democratic norms and global information integrity.
The chatbots (ChatGPT, Copilot, Gemini, Grok, DeepSeek-R1) act as conversational assistants that respond to user queries. They do not execute independent plans or make autonomous decisions in the real world, but rather generate text to assist users.
Limited Risk: The systems described are chatbots and AI-powered conversational agents, which fall under Risk Level 3 due to transparency obligations (ensuring users know they are interacting with an AI and managing generated content).
The systems evaluated are conversational LLM chatbots designed to answer user questions across a wide range of topics.
The reports provide clear, comparative empirical evidence of how five major LLMs respond to specific sensitive prompts in both English and Chinese. The findings are highly consistent across the reports, detailing the exact topics (COVID-19, Hong Kong, Tiananmen Square, Uyghurs) and the nature of the divergent responses.
The bias could be interpreted purely as a technical limitation of cross-lingual transfer rather than active data contamination or political compliance, though the report strongly points to the latter.
Technology
Flawed Multilingual Alignment
Models apply different, biased filters depending on the language of the prompt.
Model Ingestion of Web Bias
LLMs natively learn and reproduce patterns from contaminated training datasets.
Data Inputs
Contaminated Training Datasets
CCP astroturfing and state media saturate the global AI training data market.
Lack of Verified Factual Data
Developers struggle to source clean, verified datasets to train models.
Process and Methods
Inadequate Filtering Processes
Post-training filters fail to detect and remove state-sponsored propaganda.
Insufficient Post-Training Audits
Lack of robust evaluation of model outputs across different languages.
Management
Prioritizing Market Access
Firms compromise model integrity to operate data centers in mainland China.
Regulatory Environment
Strict Chinese Data Laws
PRC laws mandate that AI chatbots must 'uphold core socialist values'.
Geopolitical Compliance Pressures
Firms must comply with local censorship laws to maintain market access.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
