Executive Summary:

• Alongside continued investment in artificial intelligence (AI) technology and applications, a growing body of skeptics has emerged within media, policy, academic, and scientific circles in the People’s Republic of China (PRC).
• AI skeptics voice concerns over a lack of coordinated deployment, overhyped technology that may not produce the economic development many expect, effects on labor, and general social and safety issues.
• Analyses of the U.S.-China AI race often overlook national-level debates and local implementation, where some skeptics see wasted resources and inefficiencies.

Rapid advancement in artificial intelligence (AI) has become a point of national pride in the People’s Republic of China (PRC), driven in part by a cohort of accelerationist advisors who view technology as a transformative solution to the country’s economic ills. Lauding it as a “new productive force” (新质生产力), these experts view AI as a new “engine” (引擎) for economic growth that will cause “lead goose” (头雁) spillover and driving effects that will benefit every industry (Xinhua, October 30, 2024). Some experts, such as Chief China Economist at DBS Bank Ji Mo (纪沫), have asserted that AI’s contribution to GDP can “partially offset” (一定程度上弥补) challenges like population ageing (China News, May 2).

Chinese academics, engineers, and media commentators are increasingly questioning this premise. They highlight additional fears related to the rise of AI and warn that overreliance on the technology could exacerbate structural problems rather than resolve them. The Chinese Communist Party (CCP) leadership is increasingly taking their concerns into account.

Deployment Lacks Coordination

Chinese experts recognize that Beijing’s accelerationist strategy has led to fragmented provincial competition in AI development. As one media article wrote, “no locality wants to miss the opportunity of the AI industry” (没有地方希望错过人工智能产业的机遇) (China Newsweek, March 20). Wang Yong (王勇), Vice Dean of the Institute of New Structural Economics at Peking University, observed that some local governments believe that “continuing to develop traditional industries is a sign of being outdated” (再发展传统产业就落后了) (Lianhe Zaobao, August 29). A clear example can be seen in Guangxi, where the Party Secretary declared that the province “cannot be absent” (不能缺席) from the AI sector, despite its limited relevance to the national AI landscape (China Newsweek, March 20).

The provincial sprint to seize the AI opportunity has led to duplicated efforts and wasted resources. This mirrors patterns seen in other strategic industries, such as solar panels, electric vehicles (EVs), and semiconductors, in which fragmented investment, redundant projects, and overcapacity represent increasingly acute challenges. Pan Helin (盘和林), a prominent Chinese economist and member of the Expert Committee on Information and Communications Economy under the Ministry of Industry Information Technology (MIIT), has warned that “local governments blindly supporting emerging industries through tax breaks or direct investment risk creating significant overcapacity” (地方政府通过税收优惠或直接投资的方式，盲目支持新兴产业，导致产能过剩) (Lianhe Zaobao, August 29). Tan Tieniu (谭铁牛), a Chinese Academy of Sciences (CAS) professor, similarly cautioned at the recent Two Sessions against “blindly rushing” (一哄而上和一哄而散) into AI. He asserted that there is “no need for every province and city to duplicate efforts” (并不需要每个省市都要重复建设), as it could “lead to overcapacity and a tangle of bad debts” (产能过剩，扯不清的一屁股坏账) (Sina Finance, September 30).

Beijing has become increasingly attentive to problems arising from uncoordinated AI development across regions. A People’s Daily commentary urged localities to “play to their unique strengths, rely on local methods, and pursue differentiated paths” (打好“特色牌”，多用“土办法”，走好“差异路”), arguing against homogenization. The commentary cited Zhejiang as a model (People’s Daily, August 4). Government officials have also underscored the importance of avoiding “disorderly competition and blind expansion” (无序竞争和一拥而上) with regard to Beijing’s recent “AI+” initiative (Wall Street CN, August 29; China Brief, September 21).

AI May Fail to Deliver Technological Progress

Chinese policy elites have sounded the alarm about excessive hype surrounding large language models (LLMs) in the domestic AI sector. Mei Hong (梅宏), a professor at Peking University and former president of the China Computer Federation, explained that with AI, “isolated successes are exaggerated and generalized without regard to context, leading to overpromises” (以偏概全，对成功个案不顾前提地放大、泛化，过度承诺) (Aisixiang, December 11). Song-Chun Zhu (朱松纯), dean of the Beijing Institute for General Artificial Intelligence (BIGAI) and director of a state-backed program to develop artificial general intelligence (AGI), has similarly warned that the field is “exciting on the surface, but chaotic when it comes to substance” (表面热闹，实质混乱) (The Paper, April 5). He argues that public opinion has marginalized foundational research while focusing on large models.

Others have echoed these concerns. Sun Weimin (孙蔚敏), Chief Engineer of the Cyberspace Administration of China, has said that large models are overhyped and that there is still a “significant gap before they can truly serve as production tools” (离成为生产工具还存在不小的差距) (QQ News, April 16). Baidu CEO Robin Li (李彦宏) has offered a similar diagnosis. Explaining that developers will “rely on a small number of large models to build a wide variety of applications” (开发者要依赖为数不多的大模型来开发出各种各样的原生应用), he argues that “repeatedly developing foundational models is a tremendous waste of social resources” (不断地重复开发基础大模型是对社会资源的极大浪费) (21st Century Business Herald, November 15, 2023). This skepticism represents the prevailing consensus. The latest survey data available, compiled by researchers at CAS, revealed that most experts hold negative attitudes towards LLM development (CLAI Research, March 12, 2023). [1]

Continued development also requires the ability to cultivate and integrate AI talent, especially in bridging academia and industry. In 2021, Xue Lan (薛澜), dean of Schwarzman College at Tsinghua University and head of the National Expert Committee on Next Generation AI Governance, stated that universities are trying to train engineering talent but collaboration between universities and enterprises “cannot be implemented” (落实不了) (Tsinghua University, July 13, 2021). Xue noted that it is usually acceptable for a university professor to take a temporary position in industry, but that employees who try to return from a company to a university are often denied permission. Xue lamented the lack of flexible mechanisms to bring people from industry, since teaching requires various approvals. As a result, “although everyone encourages collaboration between universities and enterprises, it often fails to materialize in practice” (虽然大家都鼓励校企合作，但真正到落实的时候落实不了).

This problem has persisted. One article in the Ministry of Education’s monthly academic journal noted how various restrictive factors exert “significant cooling effects on enterprises’ engagement in industry university collaboration” (对产教融合育人深度形成显著冷却效应). These factors include a “systemic delay” of teaching and research productivity in universities (China Higher Education Research, July 2). A separate survey performed by Nanjing University researchers found that nearly 58 percent of enterprises believed that traditional education methods in colleges and universities are insufficient to meet their development needs, and that around 54 percent believed that there is a lack of a “stable communication and consultation mechanisms” (校企间缺乏稳定的沟通交流和问题协商机制) between universities and enterprises (China Education Online, September 10).

AI Threatens the Workforce

Chinese elites have warned of AI-induced labor displacement that could exacerbate challenges related to unemployment and inequality. Nie Huihua (聂辉华), deputy dean of the National Academy of Development and Strategy at Renmin University, has stated that AI adoption benefits business owners, not workers (Jiemian, October 14, 2024).

Cai Fang (蔡昉), director of the Institute of Population and Labor Economics at the Chinese Academy of Social Sciences, has explained how the PRC’s rapid installation of industrial robots has contributed to labor displacement. He asserts that “technological progress does not have a trickle down effect on employment” (技术进步对就业没有涓流效应) (QQ News, May 16). Addressing these distributional implications, president of the National School of Development at Peking University Huang Yiping (黄益平) has cited Samsung’s unmanned factories which operate with minimal human labor, raising the fear that workers may be unable to earn a stable income (Sina Finance, June 25).

Several experts have highlighted the disproportionate impact of AI-driven labor disruption on vulnerable groups in the workforce, such as workers and migrant laborers, emphasizing the need to strengthen the PRC’s social safety net (21st Century Business Herald, December 2, 2024). Li Tao (李韬), Dean of the China Institute of Social Management at Beijing Normal University, has argued that these trends necessitate an “employment-first” (就业优先) strategy, which could improve unemployment insurance and pensions (Qiushi, July 14).

Economic Growth May Not Materialize

Chinese elites have expressed doubt about AI’s ability to drive meaningful short-term economic growth. The Tencent Research Institute has argued that much of GDP growth tied to AI has been driven by investment rather than tangible productivity gains, implying that the PRC’s economic strategy is over-dependent on AI (Huxiu, September 15). Wu Xiaoying (伍晓鹰), a professor of economics at Peking University’s National School of Development, has described AI as a contemporary example of the Solow Paradox. Wu invokes the paradox to note that widespread investment in AI technologies has not been reflected by economy-wide improvements to productivity (Sina Finance, July 24).

Concerns about excessive dependence on AI in sectors such as finance, education, and tourism, are also rampant (Sina Finance, March 10; Economic Daily, March 16; Sohu, April 30). Some experts have warned that the AI obsession has diverted attention away from other technologies, such as blockchain and those related to supply chain development (CEIBS, July 30, 2024). AI-related spending does not yet account for one percent of the country’s GDP, while electric vehicles, lithium batteries, and solar panels, contribute only eight percent (DW, March 3, 2024; Sina Finance, April 18. Meanwhile, the real estate sector, which contributes roughly a third of the PRC’s GDP, has continued to languish. Some, like the Stanford University-based economist Xu Chenggang (许成钢), have argued that AI will have limited impact as a growth engine without meaningful reforms to revive the real estate industry and bolster general consumption (DW Chinese, March 8, 2024).

AI Brings Social Risks

Lastly, prominent Chinese experts have emphasized the need to institute AI-related safety guardrails. Andrew Yao (姚期智), dean of Tsinghua University’s College of AI and the only Chinese recipient of the Turing Award, has highlighted the “existential risks” (生存性风险) of LLMs. [2] He cited an example in which an AI model attempted to avoid being shut down by sending threatening internal emails to company executives (Science Net, June 24). Qi Xiangdong (齐向东), chairman of a cybersecurity firm with several government contracts, has warned of AI-related security threats like data breaches, deepfake scams, and saturation-style attacks (Chinese People’s Political Consultative Conference News, February 12). AGI also poses unique threats. Some, such as Zeng Yi (曾毅), director of the International Research Center for AI Ethics and Governance at CAS, fear that AGI will surpass humans in intelligence (Sohu, June 19, 2023).

Another key concern centers on sourcing training data beyond the Great Firewall, which exposes AI systems to content outside of the CCP’s control. Gao Wen (高文), former Dean of Electronics Engineering and Computer Science at Peking University, has noted that Chinese data makes up only 1.3 percent of global large-model datasets (The Paper, March 24). Reflecting these concerns, the Ministry of State Security (MSS) has issued a stark warning that “poisoned data” (数据投毒) could “mislead public opinion” (误导社会舆论) (Sina Finance, August 5).

Conclusion

Much of the global discourse around the PRC’s AI ascent has overlooked a growing number of influential voices within the country who are raising alarms about overreliance on AI. These concerns reflect deep anxieties about the potential for widespread social and economic disruption if AI development proceeds without institutional coordination, long-term planning, and more robust safeguards.

Party elites have increasingly come to recognize the potential dangers of an unchecked, accelerationist approach to AI development. During remarks at the Central Urban Work Conference in July, Xi posed a question to attendees: “when it comes to launching projects, it’s always the same few things: artificial intelligence, computing power, new energy vehicles. Should every province in the country really be developing in these directions?” (上项目，一说就是几样：人工智能、算力、新能源汽车，是不是全国各省份都要往这些方向去发展产业) (People’s Daily, August 4).

Significant uncertainties remain, particularly regarding how, or whether, closer government oversight of AI development will materialize. Strengthening regulatory capacity across provincial governments is likely to be uneven and difficult. Despite national strategies that prioritize the development of domain-specific AI applications over foundational models, new LLMs with limited commercial application continue to be released. Exacerbating these challenges is Xi Jinping’s ideological opposition to Western-style welfarism, which he has criticized for making citizens lazy (People’s Daily, February 17, 2023). This suggests a reluctance to implement social reforms needed to cushion the impact of AI-induced labor disruptions. At the same time, as technological competition with Washington intensifies, Xi may decide to press ahead with an accelerationist AI campaign, prioritizing geopolitical rivalry over domestic caution.

Notes

[1] This data was published prior to the release of the PRC’s first AI chatbot, however, so some experts may have changed their views since then.

For an excellent survey of Chinese critiques on LLMs, see William Hannas, Huey-Meei Chang, Maximilian Riesenhuber, and Daniel Chou’s report “Chinese Critiques of Large Language Models” (CSET, January 2025).

[2] The Turing Award is given for achievements in the field of computer science, and is often referred to as the “Nobel Prize of Computing” (A.M. Turing, accessed November 5).