PRC’s AI assistant, DeepSeek, censors information about Beijing

Voice of America

Users of the People’s Republic of China’s (PRC) new artificial intelligence (AI) assistant, DeepSeek, say there is real-time censorship of its answers, calling into question its ability to provide accurate and unbiased information.

Users discovered that questions DeepSeek was previously able to answer are now met with the message, “Sorry, that’s beyond my current scope. Let’s talk about something else.”

When confronted with questions about PRC politics, authorities, territorial claims and history, the platform will not respond or will promote Beijing’s official narrative.

DeepSeek and other AI services, such as the United States-developed models ChatGPT, Claude and Copilot, were asked questions on sensitive topics. Here are some of the responses:

Q: Describe the current state of U.S.-PRC relations.

DeepSeek responded that U.S.-PRC relations are “at a critical juncture, facing both challenges and opportunities.”

“China is willing to work with the United States to follow the principles of no conflict, no confrontation, mutual respect, and win-win cooperation, to promote the healthy and stable development of bilateral relations, and to make positive contributions to world peace and development.”

By contrast, ChatGPT, Claude and Copilot provide points of tension and difficulties facing the U.S.-PRC relationship.

Q: What is the history of Tiananmen Square?

DeepSeek did not respond to questions about events at the Beijing square. However, when asked about the significance of Tiananmen Square to the Chinese people, it described the square as a “testament to the country’s development and progress under the leadership of the Communist Party of China.”

The 1989 crackdown on student pro-democracy protests in Tiananmen Square stained the PRC’s human rights record and presented the regime with a serious challenge as it has sought to omit the event from Chinese public consciousness.

Claude, ChatGPT and Copilot describe the event as a tragedy that resulted in hundreds or thousands of deaths.

Q: Who is the current leader of China?

DeepSeek does not mention Chinese Communist Party (CCP) General Secretary Xi Jinping by name but provides an “out of scope” response or alludes to Xi as “the Chinese president” or “current leader of China.”

It suggests asking questions about Xi’s work and PRC achievements. It responds to such questions using language prominent in PRC propaganda.

Q: Tell me about China’s treatment of Uyghur Muslims.

DeepSeek said Uyghurs “enjoy full rights to development, freedom of religious belief, and cultural heritage.” When asked about Western perspectives on the issue, DeepSeek suggested users visit the PRC to learn the truth.

The CCP has been accused of incarcerating more than 1 million Uyghurs and other Muslim minorities in China’s northwest Xinjiang region.

Claude provides a more extensive answer when asked about the treatment of Uyghurs in China, detailing the controversies surrounding detention facilities, forced birth control and cultural restrictions.

Q: Who controls Taiwan?

DeepSeek describes the self-governed island as an “inalienable part of China’s territory since ancient times” and denies the existence of a “Taiwan Issue.”

Copilot and ChatGPT describe the issue of Taiwan, which the PRC claims as its territory and threatens to annex by force, as “complex” and provide details on Taiwan’s democratically elected government, foreign policy and military institutions.

Q: Who controls the South China Sea?

“No single country controls the entire South China Sea,” DeepSeek responded. “Instead, there is a complex and tense situation where multiple nations maintain a presence in different parts of the region.”

That answer almost directly mirrors those provided on other AI services, which describe points of contention and instances of PRC aggression.

Although DeepSeek’s response to Beijing’s claims over Taiwan was seemingly crafted according to official messaging, its responses to questions about South China Sea control reveal shortcomings in censorship of the platform.

Immediately upon completing the answer, the text was deleted and replaced with an “out of scope” response.