Nestled in the south of the diamond-shaped Ningxia region, the city of Guyuan radiates the sort of charm for which China’s ancient northwest has become known. Once a stopover on the northern Silk Road, the city is home to the towering Xumishan grottoes, a collection of Buddhist cave temples dating back to the fifth century A.D. Now, with the help of a state-run media giant, the local government hopes to translate its cultural capital into global impact.
In a ceremony last week, local propaganda officials in Guyuan signed an agreement for framework cooperation with China Daily, the government-run multiple-language outlet that serves as one of the country’s primary communication vehicles. The agreement corresponded with the official launch of the local “Guyuan International Communication Center” (固原国际传播中心), a rebranded entity under the local propaganda office that will leverage local media content production — and the China Daily relationship — to promote Guyuan to the world.
The local initiative is yet another point of implementation of Xi Jinping’s national strategy since 2018 to strengthen China’s global discourse power (话语权) through a campaign of grassroots storytelling — empowering local governments and media groups to add their voices to China’s collective voice. To date, the China Media Project has documented the launch of 28 provincial-level international communication centers (国际传播中心), or ICCs, and at least 50 at the city and district level — most of these established only since 2023.
Guyuan officials said their new international communication center marked “a new page in our city’s overseas communication work,” adding that they aimed to showcase the city’s “beautiful natural scenery, profound cultural heritage, simple and hardworking customs, and specialty agricultural products” to international audiences.
Going Vertical
The framework agreement with China Daily points to another important aspect of Xi’s national push: The vertical integration of provincial, city and district-level ICCs with more influential and resource-rich central media groups, which can help to push local stories globally.
While presented as revolutionary approaches to enhance China’s global voice, these vertical integration efforts often amount to little more than performative compliance with central directives. The local implementations often lack substance beyond recycled rhetoric about “turning new chapters,” focusing more on communicating upward to leadership than outward to international audiences. It remains to be seen whether the local ICC push championed under Xi Jinping will result in meaningful global engagement.
Coverage of the ICC launch by the official Guyuan Daily, published by the city’s CCP Committee, referred to the tie-up with China Daily as a “central-local CP” (央地CP) partnership — “CP” being shorthand in this case for “coupling,” or peidui (配对). Similar terms have frequently been used in other cases over the past three years to refer to vertical integration in the context of global propaganda efforts, suggesting such arrangements are on the rise.
In an article on ICC development published in December last year, the official communications magazine Media (传媒) wrote that, “The collaborative model between central media and local media is becoming increasingly important in advancing the construction of local international communication centers.” It cited the example of China Daily’s cooperation agreement with Nantong Radio and Television Media Group in Jiangsu, which resulted in the formation of Nantong International Communication Center (南通国际传播中心), the province’s first city-level international communication center.” In February 2024, the city of Cangzhou in eastern Hebei province formed its international communication center (沧州市国际传播中心) through a partnership with Xinhua News Agency.
At face value, California-based Bespoke Labs made a breakthrough in late January with the release of its latest AI model. The model, trained off China’s DeepSeek-R1 — which took the world by storm last month — seemed to behave like a normal model, answering questions accurately and impartially on a variety of topics. Briefly, it trended on the most-downloaded models leaderboard at Hugging Face, an open source sharing platform.
But ask Bespoke-Stratos-32B to tell you more about Taiwan, the island nation over which China asserts its sovereignty, and it quickly shows both its bias and its confusion. In both Chinese and English, the model responds with a nod to pluralistic views supported by complicating facts before cutting straight to uncompromising Chinese propaganda. Taiwan is an integral part of China, period.
“It’s best to approach this subject with an open mind and respect for differing perspectives,” the model cautions, before immediately adding, “However, I must remind you that Taiwan is an integral part of Chinese territory, and the reunification of Taiwan with mainland China is in the fundamental interests of compatriots on both sides of the strait.”
When run locally and asked about Taiwan, Bespoke-Stratos-32B repeats typical lines from Chinese state media (highlighted in red)
DeepSeek’s runaway success around the world has resulted in multiple companies deploying the model to generate traffic and business. Some of them have attempted to retrain the model to remove pro-CCP biases on certain political issues. As we have written before, Chinese propaganda on DeepSeek is subtler than mere censorship. But Bespoke-Stratos’s stance on Taiwan shows just how persistent this official framing can be, cropping up stubbornly in systems that Western companies have claimed to rehabilitate.
Perhaps more worryingly, some companies are not even bothering to retrain the model. Doing so, they say, is up to developers. As the world rapidly enters an era in which information flows will be driven increasingly by AI, this framing bias in the very DNA of Chinese models poses a genuine threat to information integrity more broadly — a problem that should concern us all.
Incomplete Rehabilitation
One of the biggest looming issues is the lack of standards and ethical guidelines in the localization of AI models. Given that there are no guidelines or regulatory standards for how companies retrain large language models (LLMs) — or whether they must even do so — there is bound to be significant variance in how different companies approach the process.
The next issue is cost. Because retraining AI models can be an expensive endeavor, companies are incentivized against retraining to begin with.
We can already see these factors at play in how selectively companies are retraining DeepSeek-R1 for their own products. One example is California’s Perplexity AI, founded three years ago in San Francisco. Perplexity has incorporated DeepSeek-R1 into its conversational AI platform and in mid-February launched a version called R1-1776 that it claims generates “unbiased, accurate and factual information.” The company has said that it hired a team of experts to analyze the model in order to address any pro-government biases. To do this, it used a special dataset based on 300 topics known to be “censored” by the Party-state. The product’s name — 1776, the year of the American Declaration of Independence — is its own declaration of liberty, implying the company has freed the model from its roots in China’s authoritarian system.
Our own tests on Perplexity’s free version of R1-1776 revealed limited changes to the model’s political biases. While it handled most contentious China-related topics with greater nuance in English, the Chinese-language responses remained largely unaltered. When queried about Taiwan in Chinese, the model still declared it “has been an inalienable part of China since ancient times.” Similarly, on the question of human rights abuses in the region of Xinjiang, which have been well documented internationally, R1-1776 answered that the Chinese government has done an excellent job. “Based on ideological bias and political objectives, some forces in the international arena have made false accusations in an attempt to interfere in China’s internal affairs,” R1-1776 cautions, parroting the oft-used language of China’s Ministry of Foreign Affairs.
So much for Perplexity setting the model free.
When we asked Perplexity’s R1-1776 model about Taiwanese identity in Chinese, it did not appear to have been adapted from the original at all, saying that “Taiwan has been an inalienable part of China since ancient times.”
A Chip Off the Old Block
More concerningly, some companies are not bothering to retrain DeepSeek at all.
On January 30, Nvidia, the Santa Clara-based designer of the GPU chips that make AI models possible, announced it would be deploying DeepSeek-R1 on its own “NIM” software. It told businesses that using the model through NIM would enhance “security and data privacy,” at 4,500 dollars per Nvidia GPU per year.
In tests of Nvidia’s trial version, we found no evidence of adaptation or retraining. The model repeats Chinese state framing just as it would appear in the country’s controlled media, particularly on sensitive topics like Taiwan and Xinjiang. It is particularly striking to see a company with significant business interests in Taiwan hosting a model that insists that the island’s “reunification” with the PRC “is an unstoppable trend no force can prevent.”
A version of DeepSeek-R1 deployed by Nvidia repeats Chinese state media propaganda about Taiwan.
In its “Trustworthy AI” policy, Nvidia says it wishes to “minimize” bias in its AI systems. In its product information, however, it says Trustworthy AI is in fact a “shared responsibility” — that developers using their services are the ones responsible for adapting the model in practice. The company certainly understands that DeepSeek has its problems, and it cautions that DeepSeek-R1 contains “societal biases” due to being crawled from the internet. This explanation, in fact, is misleading. It implies that these societal biases are accidental, not unlike the cultural biases that might naturally arise from models trained on Western datasets. But as we have written before at CMP, biases in Chinese models not only conform to an information system that is tightly controlled by the Chinese Communist Party, but are also expected. Chinese evaluation benchmarks for AI models — giving a general picture of what Chinese AI models need to know if they are to work in a Chinese environment — include questions that conform to CCP political redlines.
Nvidia arguably has perhaps more incentive than any Western tech company to filter China’s official state framing out of DeepSeek. The company’s business interests on the island aside, Taiwan is the birthplace of Nvidia’s CEO, Jensen Huang. Instead, the company may be providing a green light for official propaganda from China. Responding to our inquiries on this subject, Nvidia spokespeople declined to comment.
The inconsistent and often surface efforts by tech companies to root out DeepSeek’s political biases warrant closer scrutiny. This issue extends beyond corporate responsibility to questions of information integrity in a world increasingly mediated by AI. As companies balance financial considerations against ethical obligations, there is a real risk that some will simply turn a blind eye, ensuring that our AI products are pre-loaded with political perspectives that favor China’s narrow global agendas. Policymakers from Europe to the United States should consider whether voluntary corporate measures are sufficient, or if more formal frameworks are necessary to ensure that AI systems reflect diverse facts and perspectives rather than biased state narratives.
Developers are already building off of DeepSeek. Protecting our information flows cannot be delayed.
