by Nikola Teodosic (Policy Intern at CELI)
Policy and academic discourse in North America and Europe have focused primarily on domestic Artificial Intelligence (AI) governance frameworks, with comparatively limited attention, outside of specialized tech circles, given to China’s approach to AI development and implementation.
Public Trust in Artificial Intelligence
According to multiple studies, trust in AI in China is projected at 68% – 87%, with respondents citing comparatively high trust in AI, its benefits, while having an idea about its risks.[1][2] Furthermore, 64% of respondents from a KPMG study claimed they have formal or informal training in AI or related fields.[3] Comparatively, the level of trust in AI in the US at 32% – 41%, while roughly 50% – 58% of EU citizens would trust AI for financial matters.[4][5][6][1]
Furthermore, there is a stark difference between the perception of outsiders and mainland Chinese regarding whether they would trust China to regulate AI, with only around 27% of foreigners trusting China to do so.[7]
Chinese citizens have exceedingly high approval rates for their government’s current approach to AI development and implementation while foreigners are very skeptical of Chinese AI governance. The opposite perceptions these studies reveal suggest an inherent misconception in the West about how China approaches AI development and implementation and policymakers need a deeper understanding of why Chinese citizens are consistently positive about AI’s role in contemporary society.
Furthermore, a good start would be analyzing the recently published government “Global AI Governance Action Plan”.
The Global AI Governance Action Plan: Strategic Signaling
The Global AI Governance Action Plan released in July, 2025 is summarized in 13 relatively brief points, that proclaim China’s long-term goals and position in the world in the development and implementation of AI. Furthermore, I will highlight notable aspects from these points by comparing China’s aspirations with the current realities of their implementation, as well as how they are positioning themselves in the new space race.
Point 1. Calling for worldwide collaboration on the development and application of AI – The first point of the action plan sets a recurring theme of unity in the continuous development of AI through cooperative efforts among international organizations, states, and other stakeholders. China immediately positions itself as an enabler of internationalism in AI development.
Point 4. Supporting the Global South in developing AI technology – This point continues China’s policy of supporting AI development in developing countries which began in 2023, when a Chinese embassy argued for representation for developing countries in global AI governance.[8]
Policymakers need to be aware that China intends to expand on its broader Belt and Road Initiative and expand on (most notably) Chinese-African relations by developing new AI markets in Africa, where Chinese large language models (LLMs) would be the prime AI tools of choice. This signaling suggests that China may seek to position AI as a strategic driver of future growth and international influence, particularly in emerging markets.
Furthermore, from the perspective of raw potential, while undeniably American AI companies that run on higher-performance GPUs have higher computing power and general AI performance, DeepSeek’s energy-efficientand economically accessible advantage makes it a desirable alternative to LLMs such as ChatGPT or Gemini for the needs of the Global South’s developing AI market.
Point 5. Build and promote cross-border open-source communities and secure, reliable open-source platforms – Another point that corresponds to current online realities is that most of the top online open-source LLMs in 2025 were from China (DeepSeek-R1, Qwen 3, Kimi K2, etc.).[9] Chinese policymakers utilize this advantage to pursue advantageous international cooperation and further exposure to their dominant open-source LLMs which will continue to attract new users.
Point 7. Focus on sustainable AI – China has already proven in recent years to be a global leader in the renewable energy transition, and this approach extends to LLMs as well. DeepSeek, China’s flagship LLM, although inferior in most technical aspects to US LLMs such as OpenAI’s ChatGPT has the advantage of being cheaper and disputably up to 75% more energy-efficient to use.[10][11][12]
Point 8. Promoting common standards and norms through dialogue among standard-setting bodies of states – This point mentions the elimination of algorithm bias as a key point as well, but taking into consideration DeepSeek’s inability to even mention the Tiananmen Square massacre and, as claimed by the Wall Street Journal, mandating chatbots to pass 2,000-question ideological tests, the point raises questions regarding implementation consistency.[13] However, promoting common norms aligns with China’s AI internationalism approach.
Notably, China is trying to position itself as both a leader in international AI development and an alternative to the US, with Premier Li Qiang even proposing a new global AI cooperation organisation based in Shanghai at the city’s annual World Artificial Intelligence Conference.[14]
Point 9. Turn public services such as healthcare, education, and transportation into AI application leaders, and maintain respect for intellectual property rights such as patents and software copyrights – In April 2024, the Bejing Internet Court set a precedent in China with the country’s first “AI voice infringement case,” when it was clarifyied that: “the rights and interests of natural persons’ voices extend to AI-generated voices and defining the boundaries of rights and infringement standards.“[15]
Other countries possess similar laws protecting individuals from their rights being infringed in this manner, but this case was the first to reference the AI cloning context. Similarly, there are other regulations in place, such as the “Measures for Labeling AI-Generated and Synthesized Content” and the “Provisional Measures on the Administration of Human-Like Interactive Artificial Intelligence Services“ tackling different societal issues that arise from AI usage, such as the ease of spreading misinformation and potential addictive behavior arising from overexposure to human-like AI.[16]
Furthermore, there are separate AI integration plans specifically for branches of public administration (e.g., the National AI Healthcare Strategy).[17]
Point 9 showcases the commitment of Chinese policymakers to fully regulate AI and integrate it into every part of the economy while addressing its subsequent societal issues.
Point 12. Strengthening international cooperation on AI capacity building – This point encourages leading countries in AI to take concrete actions together to lead the education efforts for developing countries in AI application, but also for public AI literacy and skills, with specific mention being given to the interests of women and children to get accustomed to AI usage.[18][2]
Conclusion
The Chinese Global AI Governance Action Plan provides valuable insight into how China intends to use AI to expand its economic and political influence in the Global South while also promoting sustainable AI development, technology transfer, and cross-border regulatory efforts.
For policymakers outside China, the Action Plan underscores the importance of engaging with Chinese AI governance not only as a regulatory model, but as a strategic instrument of international influence, and how China‘s regulatory approach can be utilized for other international organizations and institutions that have a favourable view of AI regulation, such as the European Union.
[1] Edelman Trust Institute. “Flash poll: Trust and Artificial Intelligence at a Crossroads“, 2025. https://www.edelman.com/sites/g/files/aatuss191/files/2025-11/2025%20Edelman%20Trust%20Barometer%20Flash%20Poll%20Trust%20and%20Artificial%20Intelligence%20at%20a%20Crossroads%201.pdf Page 31. Accessed 17.01.2026.
[2] The University of Melbourne, and KPMG. “Trust, attitudes and use of Artificial Intelligence: A global study 2025“, 2025. https://kpmg.com/cn/en/home/insights/2025/05/trust-attitudes-and-use-of-artificial-intelligence-a-global-study-2025.html. Accessed 17.01.2026.
[3] Ibid.
[4] KPMG. “The American Trust in AI Paradox: Adoption Outpaces Governance“, April 29, 2025. https://kpmg.com/us/en/media/news/trust-in-ai-2025.html. Accessed 20.01.2026.
[5] Westcott, Justin. “The AI Trust Imperative: Navigating the Future with Confidence“, 2025. https://www.edelman.com/trust/2025/trust-barometer/report-tech-sector. Accessed 20.01.2026.
[6] Corduneanu, Roxana, Stacey Winters, Jan Michalski, Richard Horton, Ram Krishna Sahu, and Javier Echániz. “Europeans Are Optimistic about Generative AI but There Is More to Do to Close the Trust Gap.” Deloitte Insights, October 10, 2024. https://www.deloitte.com/us/en/insights/topics/digital-transformation/trust-in-generative-ai-in-europe.html. Accessed 20.01.2026.
[7] Poushter, Jacob. “Trust in the EU, U.S. and China to Regulate Use of AI.” Pew Research Center, October 15, 2025. https://www.pewresearch.org/global/2025/10/15/trust-in-the-eu-u-s-and-china-to-regulate-use-of-ai/. Accessed 17.01.2026.
[8] Embassy of the People’s Republic of China in Grenada. “Global AI Governance Initiative“, October 24, 2023. https://gd.china-embassy.gov.cn/eng/zxhd_1/202310/t20231024_11167412.htm. Accessed 20.01.2026.
[9] Cao, Ann. “Chinese Open-Source AI Models Occupy Top Spots among Global Developers: Ranking”. South China Morning Post, July 21, 2025. https://www.scmp.com/tech/big-tech/article/3318985/chinese-open-source-ai-models-occupy-top-spots-among-global-developers-ranking. Accessed 20.01.2026.
[10] Rinnovabili. “How energy-efficient is DeepSeek, China’s AI disruptor?“, January 29, 2025. https://www.rinnovabili.net/business/markets/deepseeks-energy-consumption-ais-75-power-cut/.
[11] MIT Technology Review also published an article greatly disputing DeepSeek’s energy efficiency. The article in question is available at: https://www.technologyreview.com/2025/01/31/1110776/deepseek-might-not-be-such-good-news-for-energy-after-all/
[12] A notable difference between the computing power of DeepSeek and ChatGPT is the graphics processing unit (GPU) that they use. DeepSeek relies (due to US export limitations on China) on Nvidia’s H20 GPU, which has noticeably lower computing power and slower in general AI performance than Nvidia’s A100 and H100 GPUs, which are explicitly focused on performance. The advantages of the H20 lie in the slightly better memory, and double memory bandwidth which improves system responsiveness.
[13] Woo, Stu. “China Is Worried AI Threatens Party Rule—and Is Trying to Tame It.” WSJ, December 23, 2025. https://www.wsj.com/tech/ai/china-is-worried-ai-threatens-party-ruleand-is-trying-to-tame-it-bfdcda2d. Accessed 20.01.2026.
[14] Goh, Brenda. “China Proposes New Global AI Cooperation Organisation”. Reuters, July 26, 2025. https://www.reuters.com/world/china/china-proposes-new-global-ai-cooperation-organisation-2025-07-26/. Accessed 20.01.2026.
[15] Huang, Xuefang, and Blair Li. “China’s First AI Voice Personality Rights Infringement Rulings and Implications.” Marks&Clerk, February 14, 2025. https://www.marks-clerk.com/insights/latest-insights/102k392-china-s-first-ai-voice-personality-rights-infringement-rulings-and-implications/. Accessed 20.01.2026.
[16] Both regulation names were translated from Mandarin using translation software.
[17] More about the National AI Healthcare Strategy is available at: https://medicalbuyer.co.in/china-launches-national-ai-healthcare-strategy/.
[18] Ministry of Foreign Affairs People’s Republic of China. “Global AI Governance Action Plan”, July 26, 2025. https://www.fmprc.gov.cn/eng./xw/zyxw/202507/t20250729_11679232.html. Accessed 20.01.2026.
