“人工智能+”——提高新质生产力的中国方案

来源：中伦视界

发布日期：2025年10月24日

近期推出的《深入实施“人工智能+”行动的指导意见》标志着中国全面入局AI。中美在AI发展目标、路径、侧重点方面差异明显。中国侧重AI垂直应用，注重民生和技术普惠，强调生态繁荣，倡导平等共享的全球AI治理。

作者丨 周洋朱熠凝 董昱辰

2025年8月21日，中国国务院发布《深入实施“人工智能+”行动的意见》（国发〔2025〕11号）（以下简称 “人工智能+” ），标志着中国积极拥抱人工智能的开放态度。

On August 21, 2025, the State Council of China (hereinafter referred to as the State Council) issued the Opinions on Further Implementing the “Artificial Intelligence Plus” Initiative (Guo Fa [2025] No.11) (hereinafter referred to as the “AI+” Initiative ), marking China’s proactive and open stance on AI.

“人工智能+”明确了中国人工智能发展的战略定位，设定了2027、2030和2035年三步走的战略步骤。在新的国际国内政治经济环境下，中国提出“人工智能+”对提高新质生产力具有战略意义。

The “AI+” Initiative declares the strategic positioning of China’s AI development and sets out a three-stage agenda for 2027, 2030, and 2035. In the new international and domestic political and economic context, China’s “AI+” Initiative is of strategic significance in boosting new quality productivity.

一、从“互联网+”到“人工智能+”，中国全面推进产业智能化

From “Internet+” to “AI+”: China’s Full-scale Promotion of Industrial Intelligence

2015年中国提出的“互联网+”行动计划，有力推动了消费互联网的繁荣发展 [1] 。电子商务、移动支付、在线出行服务、社交媒体等新消费场景不仅深刻改变了公众的生活方式，也培育了一大批平台企业。

In 2015, China launched the “Internet+” initiative, which effectively fueled the prosperity of the consumer Internet. [1] New consumption scenarios like e-commerce, mobile payment, online travel services, and social media profoundly reshaped people’s lifestyles and nurtured lots of platform enterprises.

然而，尽管“互联网+”也提到大力发展智能制造，提出要“推动互联网与制造业融合发展，发展基于互联网的个性化定制、网络化协同制造等新模式”，但其在撬动产业互联网和产业智能化发展方面作用不甚明显。中国在芯片、工业软件、企业级应用等关键技术领域与产业生态方面仍面临“卡脖子”问题。中国实现产业智能化转型亟需寻找新的突破口。

Yet, despite attempts to drive intelligent manufacturing by “promoting the integration of the Internet and manufacturing, developing new production models such as Internet-based customized production and network-based collaborative manufacturing,” the “Internet+” initiative’s boost to the development of industrial Internet and intelligent manufacturing proves insufficient. China continues to be “strangled” by strategic competitors in crucial technologies such as chips, industrial software, and enterprise-level applications. As a result, it becomes imminent for China to explore alternative pathways to advance its industrial intelligent transformation.

2022年4月，生成式人工智能大语言模型ChatGPT在美国横空出世，人工智能迅速成为全球科技引擎，对科研、技术、产业发展均带来深刻影响。2025年7月，美国白宫出台《赢得竞赛：美国人工智能行动计划》，解除人工智能监管封印，试图在人工智能竞赛中胜出并确立、维持其全球技术主导地位 [2] 。

In April 2022, OpenAI’s generative large language model ChatGPT burst onto the scene in the United States, propelling AI to the forefront as a global technological engine that has profoundly reshaped scientific research, technological innovation, and industrial development. In July 2025, the White House rolled out Winning the Race: AMERICA'S AI ACTION PLAN , a strategic move to ease regulatory constraints on AI in a bid to come out on top in the global AI race, thereby securing and maintaining global technological dominance. [2]

2023年9月，中国首次提出新质生产力概念。2025年8月，中国国务院提出“人工智能+”指导意见，明确将人工智能作为推动产业智能化发展的新质生产力，全面发展人工智能由此成为中国的国家战略。

In September 2023, China first introduced the concept of “new quality productive forces.” In August 2025, the State Council issued the “AI+” Initiative, identifying AI as a new quality productive force for fueling industrial intelligent transformation. Full advancement of AI has thus become a national strategy for China.

二、突出“普惠”与“民生”

Prioritizing Inclusivity and People’s Well-Being

1.加强哲学社会科学研究，探索形成智能向善理论体系

Strengthening Research in Philosophy and Social Sciences to Explore an “AI for Good” Theoretical Framework

“人工智能+”提出“创新哲学社会科学研究方法”、“深入研究人工智能对人类认知判断、伦理规范等方面的深层次影响和作用机理”，以及“探索形成智能向善理论体系，促进人工智能更好地造福人类”。

“AI+” Initiative calls for “innovating research methods in philosophy and social sciences,” “conducting in-depth studies on the impact and mechanisms of AI on human cognition, judgment, and ethics,” and “exploring to develop a theoretical framework for ‘Intelligence for Good’ to better serve mankind.”

可见，不论在技术层面还是从哲学意义上，中国在重视人工智能战略价值的同时，没有盲目拥抱或相信人工智能。中国对于人工智能的发展方向、发展路径、人工智能与人类的关系等基本问题仍然持探索态度。事实上，对于人工智能的原理、技术能力、方法论、社会价值、哲学意义这些问题的解答远未完成。因此，全面发展人工智能的同时，继续深入研究有关技术、哲学和社会学问题不仅十分必要，也是负责任之举。

This demonstrates that China recognizes the strategic value of AI while exercising caution, avoiding blind trust or adoption and maintaining a measured approach to both the technological and philosophical aspects. China’s exploration of AI continues, as the country seeks to address fundamental questions such as AI’s direction of development, future trajectory, and its relationship with humankind. In fact, the principles, capabilities, social values, and philosophical implications of AI are still far from settled. Therefore, it is necessary and responsible to continue in-depth research on technical, philosophical, and sociological issues while fully advancing artificial intelligence.

2.改善民生、普惠共享：是价值取向也是发展策略

Promoting People’s Well-Being and Inclusivity: A Value and a Strategic Priority

大力推动人工智能与各产业深度融合的同时，“人工智能+”在多个方面强调“民生福祉”和“技术普惠”，提出人工智能要在教育、医疗、养老等民生领域广泛投入应用。

While promoting deep integration of AI with various industries, the “AI+” Initiative also emphasizes “people’s well-being” and “technological inclusivity” in multiple dimensions. It calls for broad application of AI in public welfare sectors such as education, healthcare, and elderly care.

“民生”意味着中国力求让人工智能成为人民改善生活的公共资源，而非少数人的科技特权。换句话说，人工智能是手段，不是目的。此前外卖骑手“被困在算法中” [3] 的现象显然不符合“积极构建更有温度的智能社会”的价值取向。相较于成为剥削工具，人工智能的应用目的应为更好地造福人类。

Such highlighting of people’s well-being signals that China seeks to make artificial intelligence a public resource to improve people’s lives, i.e., a universal right accessible to all rather than a technical privilege for a select few. In essence, AI should serve as a means to achieve societal betterment, not merely an objective in itself. The previous plight of food delivery riders “trapped in the algorithm” [3] clearly runs against the vision of “building a more compassionate intelligent society.” AI should benefit mankind rather than become a tool of exploitation.

“普惠”意味着中国力求让人工智能覆盖更广泛的社会群体，让不同地区、不同年龄、不同收入的人群都能平等享受到智能化带来的福祉，而不是形成新的智能鸿沟。

Inclusivity means that China aims to extend the benefits of AI to more people, ensuring equal access to the welfare brought by intelligent technologies for people of different geographical locations, ages, and income levels, while preventing the emergence of a new technological gap.

此外，强调“民生”与“普惠”也反映出了中美在人工智能发展路径上的不同选择。在人工智能发展侧重方面，美国更强调发展超级通用人工智能，而中国则通过“发挥数据资源丰富、产业体系完备、应用场景广阔的优势”，更侧重发展对产业和民生有实用价值的垂直应用。

Moreover, the emphasis on people’s well-being and inclusivity highlights the different paths that China and the United States has respectively chosen for AI development. While the U.S. focuses on advancing general artificial superintelligence, China “leverages its strengths in abundant data resources, a complete industrial system, and abundant application scenarios” to promote practical and industry-oriented applications.

“人工智能+”发布后，交通运输部等部委颁发了《关于“人工智能+交通运输”的实施意见》，细化了推动人工智能在交通运输领域规模化创新应用的政策方向。可以预见，在“人工智能+”这一总体战略之下，农业、工业制造等领域也将有更为具体的行动方案。

After the release of the “AI+” Initiative, ministries including but not limited to the Ministry of Transport issued the Implementation Opinions on “AI+” in Transport , detailing policies to scale up AI innovation and application in the transportation sector. Under the broader “AI+” strategy, more concrete action plans are expected to be rolled out in fields such as agriculture and manufacturing.

3.中国在人工智能全球治理方面追求平等共治

China’s Pursuit of Equality and Participation in Global AI Governance

除了强调国内民生福祉和共享智能化成果，中国的“普惠”理念也具有全球维度。“人工智能+”提出要“助力各国平等参与智能化发展进程，弥合全球智能鸿沟”，同时“推动人工智能普惠共享。把人工智能作为造福人类的国际公共产品，打造平权、互信、多元、共赢的人工智能能力建设开放生态”。

China’s emphasis on inclusivity not only prioritizes domestic well-being and benefit sharing, but also transcends borders. The “AI+” Initiative emphasizes “assisting all countries in equally participating in the intelligent development process and bridging the global AI divide,” as well as “promoting AI as an international public good for humankind, fostering an open ecosystem of equality, mutual trust, diversity, and win-win cooperation.”

不同于美国追求在所谓人工智能竞赛中胜出并维持主导地位，中国在人工智能全球治理方面通过开源和生态建设，倡导多边合作和普惠共享，让技术成果惠及更多国家和群体。在当前人工智能全球治理规则仍明显缺位的背景下，平等、共治、共享的发展理念无疑提供了新的人工智能全球治理模式。

In contrast to the United States’s pursuit of dominance in the so-called “AI race”, China promotes cooperation, inclusivity, and sharing in global AI governance by building an open-source ecosystem. This enables technological achievements to benefit more countries and communities. In the absence of international rules for AI governance, China’s vision of equality, participation, and sharing certainly provides a new global AI governance model.

三、以向善价值和安全底线划定人工智能发展边界

Defining the Boundaries for AI by Benevolence and Security

“人工智能+”将“安全可控”列为人工智能发展的总体要求之一，并强调“确保人工智能发展安全、可靠、可控”，为中国未来的人工智能发展划定了边界。由此，中国人工智能发展既有了“向善”的价值方向，又有了“安全”的底线要求 [4] 。

The “AI+” Initiative identifies “safety and controllability” as one of the overarching requirements for AI development. It underscores the imperative to “ensure that AI development remains safe, reliable, and controllable.” As such, China’s AI development is guided by benevolence with strict adherence to the bottom line of security [4] .

这可能意味着在科技向善和安全底线框架内，中国将尽可能包容人工智能发展，企业在人工智能发展中充分发挥创造力，大胆探索、试错。“人工智能+”提出要“完善人工智能法律法规、伦理准则等，推进人工智能健康发展相关立法工作”。可以预见，中国在人工智能相关的法律制度、监管措施方面将迎来制度创新。

This may further suggest that China adopts a deliberately permissive approach that allows businesses to boldly innovate and experiment, learning from their mistakes during the process, as long as they adhere to benevolence and security. In parallel, the “AI+” Initiative calls for “improving AI-related laws, regulations, and ethical guidelines, and advancing the legislative process for AI’s healthy development.” Thus, it is foreseeable that China will introduce an innovative legal and regulatory regime for AI.

不同于欧盟对人工智能相对强监管，也不同于美国对人工智能的“去监管”，中国既不让监管成为创新的“紧箍咒”，也不追求无边界的人工智能竞赛，而是在向善和安全前提下鼓励人工智能与产业融合，在更多场景得到应用，最终造福人类。

Unlike the EU’s relatively strict regulatory approach, or the United States’ “deregulation” of AI, China neither allows regulation to become a “shackle” on innovation nor pursues a reckless race for AI dominance. Instead, China promotes the integration of AI with industries and its application in various scenarios under the principles of benevolence and security, so that AI can ultimately serve the well-being of humankind.

四、统筹布局算法、算力与数据资源

Overall Planning and Deployment of Algorithm, Computing Power, and Data Resources

1.全域统筹协调算力资源

Nationwide Planning and Coordination of Computing Power Resources

“人工智能+”提出“优化国家智算资源布局，完善全国一体化算力网，充分发挥‘东数西算’国家枢纽作用，加大数、算、电、网等资源协同”。

The “AI+” Initiative advocates for “optimizing the deployment of national intelligent computing resources, improving the integrated national network of computing power, giving full play to the ‘East Data, West Computing’ project as a national conduit, and enhancing synergy among data, computing, electricity, and network resources.”

“东数西算”是中国自2022年启动的一项国家战略工程 [5] ，旨在通过国家顶层设计、结合供需实际情况在全国范围内实现算力资源跨区域间调配。即利用中国西部地区在土地、气候、能源和成本等方面的优势建设算力基础设施，消化东部地区巨大的算力资源需求。

The “East Data, West Computing” project is a state-led strategic scheme launched in 2022 [5] . By virtue of a top-down national framework, the scheme seeks to align the distribution of computing power resources with actual supply and demand across regions. It seeks to leverage the advantages of western regions where land, energy, and climate conditions are more suitable for constructing data centers with lower costs, to build computing infrastructure that can accommodate the surging demand for computing power in the eastern regions.

算力资源不仅包括芯片技术，还包括通信网络和电力资源等多个方面。中国一方面通过全域统筹智能算力资源弥补人工智能芯片方面的局部劣势，另一方面“支持人工智能芯片攻坚创新”、“加快超大规模智算集群技术突破和工程落地”，以破解算力瓶颈。

Computing power resources encompass not only chip technologies but also communication networks, electricity, and other resources. On the one hand, China is working to overcome its AI chip vulnerabilities by coordinating intelligent computing resources nationwide; on the other hand, it aims to “boost breakthroughs in AI chip innovation” and “accelerate the technological advances and engineering deployment of super large-scale intelligent computing clusters” to resolve computing power bottlenecks.

因地制宜、统筹协调为提供应对算力瓶颈的短期替代方案和长期解决方案创造了可能。这一思路既回应了当前人工智能算力芯片被“卡脖子”的现实，也体现了中国“集中力量办大事”的优势。

Utilizing and coordinating local resources makes it possible for China to find both short-term alternatives and long-term solutions to computing power bottlenecks. This approach not only addresses China’s restricted access to advanced AI chips, but also leverages China’s strengths in “making great achievements by pooling resources.”

2.打通数据要素流通壁垒，适配人工智能发展

Breaking the Barriers of Data Flow to Be Compatible with AI Development

“人工智能+”倡导“以应用为导向，持续加强人工智能高质量数据集建设。完善适配人工智能发展的数据产权和版权制度”。其进一步提出“鼓励探索基于价值贡献度的数据成本补偿、收益分成方式”，同时“支持发展数据标注、数据合成等技术，培育壮大数据处理和数据服务产业”。

“AI+” Initiative calls for “development of application-oriented, high-quality datasets for AI development” and the “improvement of data property and copyright systems suitable for AI.” It further encourages exploring “data cost compensation and profit-sharing mechanisms based on value contribution.” Meanwhile，it also promotes “the development of technologies related to data labeling and data synthesis to cultivate and strengthen data processing and data service industries.”

建设高质量行业数据集和发展数据产业是实现“万物互联”到“万物智联”的智能化转型的必要条件。然而，由于数据确权制度不完善，数据还未真正成为生产要素并释放生产力潜能。

Building high-quality industry datasets and developing the data industry are essential to achieve the intelligent transformation from the Internet of Everything to the Intelligent Internet of Everything. However, due to the lack of a clearly defined data ownership regime, data has yet to become a true factor of production to unleash its full productivity potential.

为了将数据从“沉睡的资源”转化为适配人工智能发展的生产要素，需要加强数据相关法律概念与经济学概念的衔接与自洽。权利清晰、利益平衡的数据确权和流通制度将是释放数据生产要素潜能的关键。

To transform data from a “dormant resource” into a productive factor compatible with the development of AI requires the improvement of the alignment between legal and economic concepts related to data. A well-defined and balanced regime for the delineation of data ownership and data flow will be the key to unlocking the productive potential of data.

3.突出算法创新和开源生态

Promoting Algorithmic Innovation and Open-Source Ecosystems

“人工智能+”强调“健全开源贡献评价与激励机制，促进开源生态繁荣”，“发展具有国际影响力的开源项目和开发工具”。重视算法创新和开源生态体现了中国在人工智能发展路径以及全球人工智能治理方面的独特智慧。

“AI+” Initiative emphasizes “improving evaluation and incentive mechanisms for open-source contributions to foster a vibrant open-source ecosystem” and “developing internationally influential open-source projects and development tools.” Such focus on algorithmic innovation and the open-source ecosystem reflects the distinctive Chinese wisdom in AI development paths and global AI governance.

不同于美国通过算力堆叠打造“超级通用人工智能”实现赢者通吃的思路，中国则另辟蹊径。一方面通过算法创新发展“具有国际影响力的”通用人工智能，一方面通过开源和侧重垂直应用，使更多开发者可以低成本甚至免费使用中国人工智能产品开发各类垂直应用，从而形成开发者生态和行业标准。

In contrast to the U.S.’s pursuit of a winner-takes-all outcome by stacking the computing power to build “general artificial superintelligence,” China has taken an alternative path. On the one hand, China promotes algorithmic innovation to develop general AI systems “with international influence”; on the other hand, through open-source collaboration and focusing on vertical applications, China enables more developers to be able to develop diverse industry-specific applications by accessing Chinese AI products at low or even zero costs. This approach fosters a robust developer ecosystem and helps shape industry standards.

以DeepSeek为例，它没有延续美国科技巨头的算力竞赛思路，而是通过突破性的算法以Open AI二十七分之一的算力成本，达到了与顶尖模型接近的技术水准 [6] 。同时，DeepSeek开源且免费的业务模式也在全球范围内产生了生态级的影响，不仅为人工智能发展提供了新的路径，还自然形成了开发者生态和国际国内行业标准。

A standout example is DeepSeek, which did not follow the footsteps of U.S. tech giants in the computing power race. Instead, through cutting-edge algorithms, DeepSeek has achieved high technical performance comparable to leading models at only one twenty-seventh of OpenAI’s computing cost. [6] Meanwhile, DeepSeek’s open-source and free-to-use business model has generated a global ecological impact, which not only provides a new pathway for AI development but also organically fosters a developer community as well as domestic and international industry standards.

五、总结Summary

从“人工智能+”可以看出，在策略上，中国重视“从0到1”的重大科学发现，通过算法创新确保通用人工智能位居前列；同时，中国重视“从1到N”的迭代，侧重具有实用价值的垂直大模型。如此，不仅可以降低试错成本避免出现泡沫，还可以通过“小步快跑”渐进实现产业和社会智能化。此外，中国通过开源和免费，促进人工智能生态繁荣，为人工智能实现平等、共享的全球治理提供了新模式。

“AI+” Initiative reflects that, strategically, China emphasizes major “0 to 1” scientific breakthroughs to remain at the forefront of artificial general intelligence（AGI）by algorithmic innovation, while at the same time, China also values “from 1 to N” iterations, focusing on practical and application-oriented vertical models. This approach not only reduces trial-and-error costs and prevents speculative bubbles, but also progressively achieves the industrial and social intelligence by adopting an incremental approach. Furthermore, through open source and free access, China promotes a vibrant AI ecosystem and advances a model of equal and shared global AI governance.

[注]

[1]2015年第12届全国人民代表大会《政府工作报告》提出“制定‘互联网+’行动计划”。2015年7月4日，国务院印发了《关于积极推进“互联网+”行动的指导意见》。In 2015, the Report on the Work of the Government of the 12th National People’s Congress proposed to “formulate the ‘Internet +’ Action Plan.” On July 4, 2015, the State Council issued the “Guiding Opinions on Actively Promoting the ‘Internet+’ Action”.

[2]《赢得竞赛：美国人工智能行动计划》，白宫，2025年7月，第1页。 Winning the Race: AMERICA'S AI ACTION PLAN . White House, July 2025, p.1. https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf.

[3]赖祐萱：《外卖骑手、困在系统里》，载《人物》2020年第8期，第70-91页。Lai Youxuan, Food Delivery Riders, Trapped in the System, People (Renwu Magazine) , Sep. 8, 2020, p.70-71.

[4]Brian Tse, China Is Taking AI Safety Seriously. So Must the U.S ., TIME Magazine, Aug. 13, 2025, p.2.

[5]2021年5月24日，国家发展改革委等四部委联合印发《全国一体化大数据中心协同创新体系算力枢纽实施方案》，明确在8个区域打造建设大数据中心的算力枢纽节点。On May 24, 2021, the National Development and Reform Commission ( NDRC ) and other three ministries jointly issued the Implementation Plan for the Computing Power Hub Nodes of the National Integrated Big Data Center Collaborative Innovation System , designating eight regions as computing power hub nodes.

[6]Anita Kirkovska, Analysis: OpenAI o1 vs DeepSeek R1 , Vellum, https://www.vellum.ai/blog/analysis-openai-o1-vs-deepseek-r1?utm_source=bing&utm_medium=organic (last access on Oct. 12, 2025).