Sinicising AI: Zheng Yongnian on Building China’s Own Knowledge Systems
"China’s DeepSeek is a success but how it produces knowledge is no different from OpenAI’s ChatGPT and other models." - Zheng Yongnian
Beyond “AI sovereignty”, the question of how to develop a distinctly “Chinese” AI has grown increasingly prominent among Chinese scholars. Zheng Yongnian argues that to do so requires constructing local “knowledge systems” in the social sciences to ground AI in national values; otherwise, China risks sliding into Western value frameworks as the technology advances.
The idea that the social sciences provide a necessary foundation for progress in the natural sciences has a strong pedigree. During Deng Xiaoping’s first, brief return to government in 1975, a priority was rebuilding the scientific community after more than a decade of Maoist anti-intellectual tumult. Yet how was one to address the socio-political conditions that had allowed this? Realising that justifying a revived scientific culture and modernisation required theoretical backing, Deng supported the establishment of the philosophy and social science departments within the Academy of Sciences and plans for them to be separated off into a new Academy of Social Sciences. These plans came to fruition in 1977, with Deng firmly back in government.1
The dictum of “seeking truth from facts” was itself first proposed by a young philosophy student working at the Central Party School—before becoming the theoretical weapon with which Deng could legitimise his policies and delegitimise those of his rivals.2 However, according to a view increasingly prevalent today, “seeking truth from facts” thereafter developed into “seeking truth from the West”. This set the stage for a debate in the social sciences that has been further amplified over the past decade: that of universal values vs. historical particularism.
In the past decade, while Xi has extolled work in philosophy and the social sciences as the basis for Sinicised knowledge frameworks, Zhao Tingyang (of the Tianxia-system in international relations) has promoted a “China school” in IR, and the hawkish scholar Jin Canrong has called for the “decolonisation” of academic research. Meanwhile, fractures do exist. Although Yan Xuetong’s concept of “moral realism” contains a plethora of ideas drawn from traditional Chinese thought, he rejects the notion of a “China school” as motivated by political vanity rather than scientific enquiry. The economist Zhang Weiying is blunter, attributing the tendency to a “resentment complex”.
In the following interview, Zheng Yongnian holds to the scientific spirit of “seeking truth from facts”, arguing that Chinese theories should be developed in open dialogue with Western methodologies. Whether this balance can be maintained while aiming for the creation of an “autonomous knowledge system” remains to be seen.
— James Farquharson
Key Points
While drawing on Western methodologies, China should develop indigenous social theories grounded in local realities to steer technological advancement according to its core values.
Because DeepSeek’s manner of knowledge production mirrors that of Western models, relying on it merely reproduces Western intellectual “colonisation” in a localised form.
AI is accelerating an unwanted convergence of civilisations, undermining the distinct religions, ideologies and values that define them.
China can enhance the global competitiveness of its AI entrepreneurs by easing regulatory constraints; otherwise, talent and innovation will flow abroad, as exemplified by the relocation of the start-up Manus.
The pursuit of “independent knowledge systems” is fraught with challenges; too often, scholars claiming to have achieved it merely replace Western terminology with Chinese equivalents.
A true indigenous knowledge system should engage openly with global scholarship while retaining Chinese perspectives; closing off intellectual exchange or relying on official storytelling will undermine its credibility.
The K-visa programme marks a step in the right direction for attracting global talent, but further steps will inevitably be gradual and responsive to the many differing voices and interests within Chinese society.
China’s theories in politics and economics should embody its own concept of an “active society”—one intertwined with the state—rather than imitate the Western idea of a “civil society” that defines itself in opposition to it.
Rather than dividing between sycophancy and opposition, China’s think tanks should cultivate constructive, policy-focused engagement consistent with the nation’s long-term goals.
As with managing Western influence over China’s academic frameworks, the way to address tensions between the US and China in regional spheres of influence rests on “proper perception” of the realities.
Even as China consolidates its regional influence, entirely ejecting the US from the South China Sea could result in chaos; opening select facilities on Chinese islands to the US would be a practical way of acknowledging their presence.
Regarding Taiwan, the path forward is not greater urgency but “hiding one’s capabilities and biding one’s time” until a practical solution can be found; the pragmatic approach adopted by West Germany to reunification is worth consulting.
The Author
Name: Zheng Yongnian (郑永年)
Year of birth: 1962 (age: 62/63)
Position: Founding Director and X.Q. Deng Presidential Chair Professor, the Institute for International Affairs, Chinese University of Hong Kong (Shenzhen)
Formerly: Director of the East Asian Institute, National University of Singapore (2008-2019); Research director of the China Policy Institute, University of Nottingham (2005-2008); Researcher at Singapore’s East Asian Institute (1996-2005)
Research focus: International relations; Chinese politics and society; Nationalism
Education: BA-MA Peking University (1981-1988); MA-PhD Princeton University (1990-1995)
OPEN QUESTIONS | ZHENG YONGNIAN ON WHY CHINA MUST LOOK BEYOND THE WEST TO BUILD A BETTER AI
Interview by Orange Wang and Wendy Wu
Published by South China Morning Post on 15 September 2025
Many thanks to its editor for granting Sinification permission to share it here
(Illustration by OpenAI’s DALL·E 3)
SCMP: For many years, you have called for the rebuilding of China’s own knowledge system. Recently, you have also voiced concerns about “intellectual colonialism” in the artificial intelligence era. Can you elaborate?
Zheng Yongnian: The concerns mainly refer to challenges in China’s social sciences which originated from the West.
Religions, ideology, values – as Samuel Huntington explained in his book The Clash of Civilisations – are important for any nation. And the meaning of society and technology is determined by the humanities and social sciences.
Chinese researchers learned and adopted theories of Western social sciences, but these are based on Western methods that summarise Western practices and experiences, and are then used to explain Western society.
Those theories have failed to explain Confucian civilisation, the Islamic world and Indian society. We should fully embrace our secular civilisation, and thereby play a proper role in the international order.
In the AI era, knowledge production and distribution have picked up pace. But the logic of using AI during the process remains unchanged: the feeding in of information and its distillation.
China’s DeepSeek is a success but how it produces knowledge is no different from OpenAI’s ChatGPT and other models.
In the era of AI, knowledge production has become more a process of convergence instead of divergence, which puts us at greater risk of relying on the Western knowledge system.
SCMP: Does this mean that China’s academic and strategic communities need to be especially cautious when using AI tools? Where should we begin such a massive task as building a governance philosophy and knowledge system tailored to China’s needs? How do we learn from the West while avoiding dependence on Western frameworks?
Zheng Yongnian: It is very dangerous that many people are using tools like ChatGPT without any awareness of the problems. The issue lies in the “feeding” of the model.
In the natural sciences, since modern times, all the best papers have been published in foreign languages, especially English. Even Chinese scientists and engineers publish their best work in English, as the top journals recognised by Chinese scholars are predominantly Western. The same is true in the social sciences.
It is now very hard to avoid a process of colonisation if you want to use ChatGPT, DeepSeek or similar tools to build a knowledge system. We need to look at the source and reshape the feeding aspect.
China’s thousands of years-long uninterrupted civilisation offers vast practical experience and historical records unmatched by others. But transforming those into modern social science requires dedicated effort.
China has gone through massive transformations in modern times, not to mention 40 years of reform and opening up. Those efforts spanning generations should be harnessed to prepare the feed.
Otherwise, it remains one-way, with only Western inputs and none from China. If we built equally strong Chinese feeds, it would be more balanced.
I am not rejecting the West. The problem is that the social sciences lack genuine Chinese “stories”. Even when we do tell a Chinese story, the theoretical framework is still Western. China merely provides evidence – and often the corroboration is incorrect, because the proposition itself is flawed.
Even more dangerous is the short-sightedness of many Chinese scholars. Many universities and research institutes claim to have already built an “independent knowledge system”. But they are talking to themselves.
Some economics books simply replace “Western” in the title with “Chinese”—this is not an independent knowledge system.
We need to stay grounded and take a truth-seeking approach. It will be impossible to accomplish the work in one or two years, or with one or two persons.
SCMP: You mentioned that building an independent Chinese knowledge system should start from practice. Which of China’s own practices do you think deserves particular attention?
Zheng Yongnian: The Sinicisation of Marxism during Mao Zedong’s era is an example – it was grounded in practice.
At that time, European revolutionary theory emphasised urban struggle. Some failed experiences led Mao to re-examine orthodox Marxism and innovate. He raised “encircling the cities from the countryside”, which started a revolution in rural areas.
Deng Xiaoping was also innovative, with examples such as the “socialist market economy” and the “cat theory”.
We have many innovative practices today but they have not been developed into conceptual theories. Instead, we often use Western theories to interpret our own practices, but that does not work. They are fundamentally different things.
There are also many practices worthy of attention in society, such as the bottom-up stories behind Shenzhen’s development.
Justin Lin Yifu proposed new structural economics, highlighting a well-functioning government and an efficient market. I think both ideas already exist in Western economics [Note: We covered Lin’s theories and Zhang Weiying’s criticisms of them in an earlier two-part piece]. What’s missing, in my view, is a third element: an “active society”.
“Civil society” is a Western concept which implies opposition to the state. But Chinese society has never been like that.
From state to society, there is no clear boundary—it’s a continuum. Our market and state are not strictly separate either. We don’t have the Western concepts of absolute private or absolute public ownership.
Ancient Chinese were already experimenting with the well-field system two millennia ago, showing that our ancients sought to combine public and private ownership, which cannot be appreciated through the Western framework.
SCMP: You also said it was important to break the West’s intellectual monopoly. In today’s tense China-US climate, some may see this as a move by China to compete for global leadership and narrative power. How do you view it?
Zheng Yongnian: That is about viewing things through an ideological lens. But ideology is not something scholars can control – we can only feel helpless about it.
I have always advocated that we are not rejecting the West at all. We need to learn from Western methodologies, which are very important. But the concepts and theories generated by those tools often do not fit China’s reality.
China should take their methods, while setting aside their theoretical constructs, to re-examine China’s own practices and develop its own concepts and theories, just as each Western country has its own distinct social science characteristics.
If Chinese scholars can build a Chinese social science, it would be a tremendous contribution to global social sciences – but it would not be an act of confrontation.
Some so-called China experts are distorting things, but those who see the issue from an ideological perspective are a minority, in my view.
The mainstream has seen the reality of China’s development and transformation but cannot explain it with Western frameworks. They are waiting for an answer. If we do not provide it, who will?
That is not about shutting ourselves off and opposing everything Western. A pile of self-talk will not work—nobody would understand or accept it.
Building a knowledge system is about communication. We accepted Western theories not because they were forcibly imposed but because we found their logic convincing. China should create something that can work in the same way.
SCMP: How do you expect the rebuilding of China’s knowledge system in the age of AI to feature in the 15th five-year plan?
Zheng Yongnian: I think the effort to build a Chinese knowledge system will accelerate, since the urgency is already here.
We often say the West does not understand us, misunderstands us or even demonises us. It is now our responsibility to interpret ourselves to the West. With China’s role on the global stage becoming increasingly significant, we cannot keep others waiting for us to do so.
We also should be more open. The Western knowledge system was built through openness and we must do the same. A genuine knowledge system must be intelligible to everyone.
Meanwhile, we often pretend to have understood the West, but effectively we only grasp the surface. We truly need to cultivate a large pool of talent who are proficient in various civilisations.
I think the current situation is more challenging than during the Republican era.
Why was Fei Xiaotong able to write Peasant Life in China? He used Western methodologies to interpret Chinese phenomena brilliantly, which can be seen as an element of China’s independent knowledge system.
Chinese studies and Western studies are now separated, whereas during the Republican era, they were integrated.
Current academic training focuses heavily on the quantifiable and measurable—the very things ChatGPT can do—but the truly valuable aspects of the social sciences and humanities are those which cannot be measured.
Still, China also enjoys a major advantage: our rich practical experience.
SCMP: In your latest book Reconstructing Chinese Knowledge in the AI Era, you raise the need to avoid an excessive focus on “political correctness” and “ideology” in social science research. How can that be achieved? Why have you called for a “de-witching” of the intellectual community, your term for a witch-hunting mindset?
Zheng Yongnian: Many humanities scholars today are more like secular priests, chasing popularity by producing endless “chicken soup for the soul” content which has been rampant in China. That is of little use to society and is truly not good.
If, while the West uses an ideological perspective to demonise China, Chinese scholars also engage in using ideology to demonise the West, that is not building a knowledge system, it is just tit-for-tat name-calling.
That adds no value and is fake scholarship, which may be catchy but will leave nothing lasting. We need to be self-aware. We must return to basic facts, to science and to rationality.
SCMP: How can China’s think tanks be developed to respond effectively to international opinion and increase their influence? You have spoken positively about the Rand Corporation in the US. What lessons could Chinese institutions learn from it?
Zheng Yongnian: The US has the most developed think tanks in the West and, in terms of serving national interests, I think the Rand Corporation does the best job.
Chinese think tanks can be roughly divided into two categories: the “justification” type—that just approves of whatever the government asks it to look at—and the purely critical, that says everything the government does is wrong. Both inevitably have their place.
Every government needs someone to assess whether a policy is scientifically sound. But if you can never say it is not, that is not really a think tank. Likewise, if it simply claims every government action is no good, that is not a think tank either.
A true think tank should serve national interests. All leaders are also meant to serve national interests. For example, if a Hong Kong think tank served the chief executive personally, that could be problematic. If it served the interests of Hong Kong, that would be good.
SCMP: That brings to mind the issue of soft power. How can China build a knowledge system that is both independent and universally applicable?
Zheng Yongnian: I think there are four conditions when it comes to building soft power.
First, what you write must be read by someone. If no one wants to read it, it will not work. Second, people must be able to understand it. Third, after understanding it, people accept it. Fourth, not only do people accept it, they also spread it to others.
We currently have problems in all four stages. The West, however, has succeeded in some of those areas.
Our teachers have been spreading Western knowledge, but when will we be able to write books that reach Western classrooms? We are still far from that. It may be happening in the science and engineering fields, but what about the social sciences?
Take China’s experience of lifting 800 million people out of poverty in four decades – which is also recognised in the West as a miracle in global economic history. How do we tell the story of poverty alleviation?
I think we should use academic language. If we only emphasise ideological correctness, that certainly will not show the full picture. Telling the story in political or ideological language also bars others from learning from us.
SCMP: Beyond its influence on social structures, do you expect AI to further deepen the gaps and tensions between nations? Could that have an impact on international relations?
Zheng Yongnian: Of course.
Both domestic and international political-economic orders need a base, just as in the previous three industrial revolutions. In the AI era, artificial intelligence is becoming the base, and both domestic and international orders will be different.
I am somewhat pessimistic.
The first three industrial revolutions liberated humans from physical labour. But AI is freeing humans from intellectual labour, which means many people may become unable to think. That leads to what I call “artificial ignorance”, a self-inflicted intellectual harm.
There is a convergence trend between China and the US in AI development and, especially under open-source conditions, content will look increasingly similar in the future.
When it comes to AI, given the American tech right, the US is following a development-oriented model, with no current regulatory framework at the federal level.
Meanwhile, the EU is over-regulating, which has stifled innovation—a very sad story.
Some European leaders have called for deregulation to boost indigenous AI companies, but that will lead to a dilemma, since simultaneously strengthening regulations on US firms will be a mission impossible with the EU commitment to consistency in rules.
China now sits somewhere between the US and the EU, with some overregulation. This was reflected by the recent relocation of Wuhan’s AI start-up Manus to Singapore.
One particularly important aspect of US President Donald Trump’s reforms has been deregulating business – the technological accelerationism advocated by the tech right.
If the US deregulates while the EU, China and India do not, emerging technologies within our country that cannot be adopted on the ground due to excessive regulation at home are likely to flow to the US—a phenomenon which has already been happening for years.
China should ensure proper regulation on environmental protection and labour rights, but in terms of getting technologies into application, I think regulations should be relaxed.
Otherwise, the much-discussed “new productive forces” will not be able to materialise and transform into real economic activities, keeping China in intense “involution”. If new economic activity were more abundant, there would be less involution.
SCMP: What changes in the Ukraine war can we expect after the Putin-Trump Alaska summit and how will they affect China? Also, given the tensions between India and Washington, is there a possible tilt in the balance of the US-China-India triangle?
Zheng Yongnian: The prospects for resolving the Ukraine war are full of uncertainties. Neither side has shown signals of compromise. Europe won’t accept a land-for-peace swap plan, which it might deem a nightmare.
China is indeed the outsider. It has been held hostage by the war, yet is completely shut out when efforts are needed to work out a solution to end the war.
There are two images of China’s strength. One is measurable – such as GDP and defence progress—when China moves towards the centre of the global stage.
On the other hand—and this brings us back to the matter of an indigenous and independent knowledge system—it is about how we identify ourselves and how others recognise us. It is not only about strategic thinking, it is also a philosophical issue.
There has never been a fixed triangle relationship, regardless of the major ones between the US, China and Russia, or US-China-India ties. Despite signs of decline, the US is still the global hegemony.
India is taking an opportunistic approach. Despite recent tensions with Trump, India and the US have moved closer, but India will never be a true ally for the US. India is also pursuing regional hegemony, in particular, it has tried to set up its own Global South group to exclude China.
We still have not fully understood or grasped the Western language of international relations—or the multilateralism and alliances of the Western narrative. In China, what we call a partnership or strategic partnership is different from what the US terms an alliance, for example.
Neighbourhood diplomacy matters as well. China’s rise in the world cannot be realised without its rise in Asia. As Mao Zedong said, it is very important to identify who is friend and who is enemy. Treating everyone as a friend often means having no genuine friends.
SCMP: Tensions between Beijing and Manila have been rising. The Philippines has also tightened security with the US and its allies. You have said Beijing needs to maintain strategic patience in the South China Sea. Does the recent collision between Chinese vessels during a confrontation with the Philippines near Scarborough Shoal mean it is time for a different approach?
Zheng Yongnian: The interests of the US are deeply embedded globally, such as in Europe, the Middle East and Asia. But we need to accept the fact that the US cannot be driven away, nor can it successfully or completely retreat from those regions.
The prerequisite to resolving the South China Sea disputes is to acknowledge the US presence in the region, and we should also be aware that the ultimate goal is not to expel its influence from the Western Pacific.
When the US withdraws from a region, chaos often follows – as we have witnessed in the Middle East and Ukraine. The affected countries often end up asking the US to return.
We may need to think out of the box and break with orthodox thinking. China is a strong advocate for globalisation and it might not be in line with our push to overemphasise external forces. Do we need to go to Latin America? Do we need a European market? Of course—we are all intertwined.
Our perception of facts depends on our mindset. In my view, the South China Sea dispute is not a big headache. What matters more is how we understand the issue. Without a proper perception, it will be hard to build mutual trust.
If both sides deemed the removal of the US out of the Western Pacific was the final goal, that might be the worst scenario.
Just as we used to allow US warships to make port calls in Hong Kong, we could consider the possibility of opening certain infrastructure on islands in the South China Sea to American vessels as an international public good.
That might allow more opportunities to solve the issue fundamentally.
SCMP: The South China Sea disputes are increasingly complicated by the Taiwan issue. How can the situation be managed to avoid miscalculation and mitigate risks?
Zheng Yongnian: We still need strategic patience.
Deng Xiaoping raised the strategy of “hiding capabilities and biding time”—which many in the West interpret as waiting to take revenge. That is not the meaning. It refers to seeking a rational, better solution while exercising patience. As Premier Li Qiang once said, there are more solutions than problems.
We need to seek truth from facts and learn from history—such as West Germany, which offered citizenship to East Germans, offered exchange rate parity between the two currencies and levied a 5.5 per cent unity tax on West Germans.
We won’t drop the use of the military as a last resort and we should prevent any attempt at independence. But there is much room for us on social and economic fronts to move towards reunification and attract Taiwanese people.
We have talked a lot about the political will of reunification and the issues from ideological perspectives, but we seldom think about how to achieve it in concrete moves. It is time to think.
SCMP: Is China’s new K visa for STEM professionals likely to be effective in attracting foreign talent to China?
Zheng Yongnian: Visa policy matters. The K visa is only the first step. From a policy standpoint, it may be China’s first visa specifically designed for talent. Look at the US – it has various visa categories for different types of talent.
But once someone gets a K visa, you still have to consider other issues, such as how to arrange things for the person’s family. Attracting talent is not just about a visa or money. It is about the ecosystem, on which China still has a lot of work to do.
It is moving in the right direction. China will definitely become more open in education and talent attraction.
There is likely to be a series of new measures in the higher education area—including in relation to the operation of foreign universities in China—rolled out during the 15th five-year plan period.
China is a large country and different social groups have varying views on certain issues. Advancing step by step has its advantages, but the direction must be maintained. We should continue moving steadily forward.
SCMP: Does the growing international popularity of Chinese dramas, games and toys suggest that relaxing regulations could also help to tell China’s stories well?
Zheng Yongnian: Yes. China’s stories can take many forms, including written forms and those you just mentioned. And the relaxation of regulations is needed.
We cannot see the West as homogeneous. It is made up of diverse social groups with different needs. That means we need more people, using different ways, to tell China’s stories.
We really need an active society. If you rely solely on a government-commissioned institution to tell China’s stories, it will certainly not succeed.
READ MORE
China’s AI Path and the Needham Question: From 1 to 10, Not 0 to 1
As both nations invest ever more significant resources into artificial intelligence, the divergent trajectories between China and the US have become a key issue—with many pointing to a philosophical—even quasi-religious—divide between the rapid adoption of practical AI applications in the former and the more abstract pursuit of AGI in the latter. Huang Ping, an up-and-coming associate professor at the Chinese University of Hong Kong, agrees with these analyses. He notes that despite state-funded investment drives in basic research, China’s innovation system remains relatively weak at original breakthroughs (moving from “0 to 1”) but excels at scaling and commercialising technologies (moving from “1 to 10”).
Western Ideological Exhaustion and China's Trump Opportunity by Zheng Yongnian
Within China, Zheng Yongnian is sometimes viewed as a ‘Western’ academic. Trained in the US, he has long advocated greater opening-up of China’s economy and has expressed scepticism towards claims portraying China’s model as an ideologically superior alternative. His polemical assertion—that the West has not produced any meaningful political ideas since...
Ezra Vogel, Deng Xiaoping and the Transformation of China (Harvard University Press, 2013), 133, 205.
Ezra Vogel, Deng Xiaoping and the Transformation of China (Harvard University Press, 2013), 211.
 | A guest post by
|