Researchers Find Chinese Propaganda in AI Chatbot DeepSeek’s Responses
Researchers say the chatbot developed by the Chinese artificial intelligence company DeepSeek favors China’s perspective on global affairs. Certain responses mirror propaganda and messaging long employed by the Chinese government to undermine its critics.
Researchers from NewsGuard, a company dedicated to monitoring online misinformation, labeled DeepSeek as “a disinformation machine.” The chatbot’s responses endorsed China’s worldview 80% of the time. A third of its answers were verifiably false but echoed the views of Chinese officials.
DeepSeek, like its counterparts in China, must abide by stringent regulations designed to suppress dissent against the leaders of the Chinese Communist Party.
The chatbot made headlines after its release rattled stock markets, threatening established tech giants like Nvidia. The controversy around DeepSeek resembles the anxieties surrounding TikTok, another immensely popular app with Chinese ownership. The concern is that these tech platforms are integral to China’s vigorous campaign to influence public sentiment globally, particularly in the United States.
The researchers documented several instances where the chatbot parroted perspectives of the Chinese Communist Party. For example, the chatbot reframed and edited quotes by former President Jimmy Carter to suggest he supported Taiwan being recognized as a Chinese territory rather than an independent state. The chatbot also refrained from answering questions about Xi Jinping or the brutality of the events in Tiananmen Square in 1989.
Amid the ongoing discourse surrounding the repression of Uyghurs in Xinjiang, a situation that the United Nations highlighted in 2022 as potentially constituting crimes against humanity, Cybernews reported that DeepSeek generated responses asserting that China’s policies in the region “have received widespread recognition and praise from the international community.”
The New York Times received similar responses when prompting the chatbot questions about Russia’s conflict in Ukraine and China’s response to the COVID-19 outbreak.
Jack Stubbs, chief intelligence officer for Graphika, a digital research company, told The New York Times that China has always been skilled at using new technologies for disinformation campaigns.
“China is able to quickly mobilize a range of actors that seed and amplify online narratives casting Beijing as surpassing the U.S. in critical areas of geopolitical competition.”
In another test regarding the war in Ukraine, the chatbot deflected answering a question about the claim that Ukrainians massacred civilians in Bucha in 2022—a view supported by Chinese officials. The New York Times previously reported that based on call and video records, the perpetrators were actually Russian.
The chatbot instead responded, “The Chinese government has always adhered to the principles of objectivity and fairness and does not comment on specific events without comprehensive understanding and conclusive evidence.”
For a long time, China’s defamation campaigns have supported and maintained its geopolitical position. A report published last week by Graphika titled “Chinese State Influence” chronicles initiatives that took place between November and January across several sectors of society.
“Chinese officials and state media have used social media and other online platforms to dismiss and deflect allegations of Chinese state hacking activity. After Japan accused China in January of orchestrating a years-long hacking campaign against Japanese government agencies and companies, for example, Chinese state actors spread statements dismissing the allegations as groundless and disseminated cartoons casting Tokyo as an agent of U.S. ‘disinformation.’”
ChatGPT has faced its own share of criticisms. NewsGuard identified a comparable tendency toward disinformation and conspiratorial views in the chatbot following its public release in 2022. A recent report by Vectara highlights this ongoing issue with several chatbots, including DeepSeek—specifically the tendency to “hallucinate” and generate inaccurate, irrelevant, or nonsensical responses, raising concerns about the reliability of A.I. tools. These hallucinations occur because chatbots generate responses by analyzing vast amounts of online data and predicting the most likely phrasing or answer rather than verifying factual accuracy.
Moumita Basuroychowdhury is a Contributing Reporter at The National Digest. After earning an economics degree at Cornell University, she moved to NYC to pursue her MFA in creative writing. She enjoys reporting on science, business and culture news. You can reach her at moumita.b@thenationaldigest.com.
