Liu Wei is Director of the Human-Computer Interaction and Cognitive Engineering Laboratory at Beijing University of Posts and Telecommunications.

The global AI industry has entered an unprecedented phase of rapid expansion. According to the International Data Corporation (IDC), the market size for artificial intelligence (AI) reached $638.23 billion in 2024, marking an impressive year-on-year growth rate of 21.5 percent.

 

While the United States continues to lead globally in AI financing and research, commanding around 70 percent of the world’s private investment in the sector, breakthroughs from countries like China and others are gradually narrowing this gap, especially in the development of large language models and AI applications.

Chinese President Xi Jinping during an April 2025 visit to the Shanghai Foundation Model Innovation Center | Source: XinhuaNet; english.news.cn

 

Once an underdog in the global AI race, China is now hurtling to the front of the pack to assume a leading role in AI development. On the technological front, the country has made significant strides in developing open-weight reasoning models. A key marker of this progress is DeepSeek-R1, a Chinese open-source model whose performance rivals that of OpenAI’s o1. China’s efforts to catch up with and even surpass established AI leaders have been marked by notable technological advancements, particularly in the development of open-weight reasoning models, of which DeepSeek-R1 is a prime example.

 

In terms of industrial growth, China’s core AI sector, now valued at nearly 600 billion yuan, is projected to grow at a compound annual rate of 15.6 percent over the next decade; by 2030, its market size is expected to surpass 1 trillion yuan.

 

A highlight of China’s AI drive are its rapid advances in embodied intelligence and humanoid robotics, where Unitree Robotics, a tech company based in Zhejiang Province’s city of Hangzhou, currently leads the charge with world-class motion control and environmental perception technologies.

 

These achievements have been largely driven by targeted policies and solid financial support from the Chinese government, whose push for the AI Plus initiative and the construction of high-tech industrial parks nationwide are accelerating the shift to smart factories and intelligent agriculture, as well as the application of AI technologies in health, education, finance, and other industries in the service sector.

 

Take DeepSeek as an example. The company’s flagship reasoning model has already been deployed in 228 top-tier hospitals across the country to assist with a wide range of services, including clinical decision support, patient record tracking, and medical imaging.

 

Meanwhile, in response to U.S. export controls on advanced computing and semiconductors to China, Chinese enterprises are ramping up efforts to develop in-house AI solutions. This move is part of a broader effort to reduce reliance on foreign-made AI chips. Hopes are high that the country will continue to push forward in the global AI race with greater strength and ever more innovative solutions.

 

The Growth Trajectory of China’s AI Sector

The evolution of China’s AI industry can roughly be broken down into three stages. Each is examined in turn.

 

1. The Exploration Phase (1979–2012)
   During this early phase, the focus was on laying the theoretical groundwork for AI development. Research was the primary objective, and China invested heavily in building a foundation of technological expertise, which would later prove essential for more applied AI breakthroughs.

 

2. The Breakthrough Phase (2012–2019)
   This phase marked a shift from theoretical research to practical implementation, as AI innovations in computer vision, natural language processing, voice recognition, and customer service chatbots began to proliferate across various industries.

 

3. The Industrialization Phase (2020–present)
   With ongoing upgrades to AI infrastructure and an ever-wider deployment of the technology, the sector has entered a period of rapid industrial expansion. As of now, the country has established a highly integrated AI ecosystem spanning more than 4,500 enterprises and covering every link in the supply chain, from chips and algorithms to data, digital platforms, and real-world applications. The domestic AI market is already valued at nearly 600 billion yuan and is poised for continued expansion as China further solidifies its position as a global leader in AI development.

 

Several key advantages have underpinned China’s ascent through each stage of its AI evolution:

 

1. Sheer market scale
   With its massive consumer base, China provides an ideal environment for testing and deploying AI solutions. The sheer size of the Chinese market facilitates the widespread adoption of AI across multiple industries.

 

2. Robust policy backing
   The Chinese government has introduced a stream of stimulative policies to accelerate AI research and application and support the sustainable development of the AI industry.

 

3. A solid talent pool
   Home to the world’s largest higher education system, China has a vast reserve of tech talent, with 45 percent of university graduates specializing in fields like science, technology, engineering, and mathematics. This large pool of talent ensures a steady supply of skilled workers for the country’s AI sector.

 

4. Rich data resources
   As the world’s largest manufacturing hub, China generates immense volumes of industrial data, which serve as invaluable assets for training AI models.

 

5. Strong adaptability
   Supported by the world’s largest scientific and research network, China leads globally in AI-focused academic publications, citations, and patent applications.

 

6. A complete industrial chain
   China’s extensive manufacturing base and comprehensive supply chains in electronics and machinery have enabled the rapid development and deployment of AI technologies.

 

Together, these factors have positioned China as a powerhouse in global AI development.

 

The Human-Machine-Environment System
As AI sets out to permeate almost every aspect of our daily lives, humans and machines are becoming ever more deeply intertwined, bringing forth a host of new challenges: how can we build AI systems that are not only smart, but also transparent, trustworthy, and consistently cooperative? And how do we avoid becoming overly dependent on AI and relinquishing our own cognitive and practical abilities?

 

A major trend that has emerged in response to these challenges is the human-machine-environment system, an approach to AI development that envisions an integrated system in which each element—humans, machines, and the environment—works together in a highly efficient and coordinated manner.

 

Under this model, humans are the users and controllers, interacting with AI through a multitude of channels, including natural language and visual interfaces; machines, including computers, robots, and other smart devices, interpret human input and environmental data to provide intelligent responses; the environment refers to both the physical context (such as smart buildings and intelligent urban transport systems) and virtual space (such as cloud computing services and big data) in which human-machine interaction occurs.

 

This solution is not about replacing humans with machines; it is about enabling each to do what they do best. While machines excel in rapid processing, precise execution, and data analysis, humans are unsurpassable in their creativity, empathy, and complex decisionmaking. When these capabilities are effectively combined, the result is not just greater efficiency but also greater intelligence.

 

This requires AI not only to understand user needs but to adapt to environmental changes by proactively adjusting its responses. In a smart home, AI-powered devices can automatically adjust lighting, temperature, and other settings based on user behavior, environmental conditions, and energy use; a smart traffic system can greatly reduce congestion and improve road safety by adjusting traffic light durations according to real-time traffic flow; an autonomous vehicle can continuously scan and respond to its changing surroundings, thereby ensuring safe navigation.

 

An integrated human-machine intelligence system also allows users to interact with machines in multiple ways, such as through voice commands and hand gestures.

 

These scenarios depend on the deep integration of multiple disciplines, including but not limited to computer science, psychology, and sociology. Therefore, interdisciplinary cooperation should be strengthened to further drive the development of human-machine systems.

 

This kind of system also marks a shift toward a more user-centered approach to AI development. However, the field also raises multiple ethical and sustainability concerns. Therefore, it is essential to ensure that the development of powerful digital minds serves the common good, and that unintended social harms are kept at bay.

 

The Social Impacts of AI
AI’s role as a new factor of production has significant implications for the global economy. By improving productivity and optimizing resource allocation, AI can foster greater economic growth. In finance, for example, AI’s ability to process vast amounts of data enables investment firms to optimize portfolios, refine strategies, and improve returns, ultimately leading to more efficient capital flows and faster economic development.

 

Furthermore, AI can foster the growth of emerging industries and open up new space for expansion. An example would be smart homes, an emerging sector where the rise of AI has led to the creation of a plethora of intelligent devices, such as smart speakers and locks.

 

Technology can also revive traditional industries by facilitating their digital upgrades, improving productivity, reducing costs, and enabling a rise along the global supply chain, which ultimately can enhance the overall economic strength of a country. In the manufacturing sector, for instance, the use of AI-based vision inspection technology can ensure both product quality and a more streamlined production flow by improving the accuracy and speed of quality tests and reducing the risks of defects.

 

However, benefits also come with side effects. For instance, the automation of low-skilled jobs, such as those in customer service, warehouse management, and production line work, is reducing employment opportunities for certain groups of workers. Meanwhile, competition for jobs that require a high level of professional skills, such as accountants and legal assistants, is also intensifying, as these roles may likewise be replaced by AI.

 

AI-powered systems, which often rely on vast amounts of user data for tailored services and personalized recommendations, also raise serious privacy and security concerns. If this data is misused, it could pose significant threats to individual privacy and data safety. The widespread use of smart surveillance cameras and facial recognition technology means that all aspects of our everyday lives could be monitored. The ubiquity of such surveillance not only risks violating personal privacy but may also lead to “information overload.”

 

The lack of transparency in AI-powered decisionmaking entails many ethical risks. For instance, the application of AI to fields like justice and job recruitment has sparked worries over fairness and accountability. The fact that AI models are trained on large datasets means that they can reinforce existing societal prejudices and perpetuate biases.

 

The rise of AI may also widen social divides, as only a very limited number of tech titans may eventually benefit from the AI rush. These enterprises have dominated the AI scene with an unchallenged lead in technological advancements and massive investment in research and development, while smaller businesses and low-income groups risk being completely marginalized. When it comes to education, AI may widen the divide between children from the upper class, who have a better chance of enjoying quality AI-powered educational resources, and those from poorer families, who often lack access to Internet resources.

 

Last but not least, current AI models do not yet possess the ability to distinguish truth from falsehood. Nor can they recognize their own limitations. The content these models generate is more akin to a reconstruction of existing data. (GLOSA) The intrinsically limited scope of their data samples means that accuracy can never be fully guaranteed. The unpredictable nature of complex AI models also adds another layer of instability and insecurity.

 

Addressing AI’s double-edged nature requires proactive measures. Strengthening regulation and oversight of AI use is crucial to ensuring its safety and reliability; by improving the transparency of AI reasoning models, users will better understand how AI arrives at conclusions and thereby build deeper trust in these digital tools; more efforts should be invested in creating policies and providing social support to ensure that the deployment of AI is equitable and widely beneficial to everyone; clear ethical and legal frameworks must be established to guide AI development so that it always aligns with social values. Additionally, promoting the integration of AI with traditional industries and fostering the growth of a skilled AI workforce will help mitigate the many negative impacts it may produce on employment.

 

Global AI Governance
The state of global cooperation on AI governance is one of both promise and mounting challenge. Governments and organizations worldwide have begun to grasp the profound influence AI may have on innovation, economic growth and democracy. As a result, considerable resources have been directed towards creating relevant policies and investing in AI research and development.

 

The United States, for instance, has launched the National Artificial Intelligence Initiative Office, which serves as the central hub for federal coordination and collaboration in AI research and policymaking across all sectors. Several proposals made by the U.S. National Security Council on integrating AI into the country’s national strategy have also been adopted, and numerous legislations passed on better regulating the use of AI.

 

Globally, more than 60 countries and regions, along with the European Union, have launched over 700 AI-related initiatives, forming a broad push to advance and govern AI.

 

However, there are still many challenges to be tackled. The global digital divide persists, with only seven of the 193 UN member states active in global AI governance, while 118 countries, mainly those from the Global South, have been notably absent from the scene. Furthermore, countries have varying standards for AI governance, which often conflict with one another. Together, these unique risks posed by AI highlight the urgent need for a universally accepted framework for regulating AI development.

 

To address these challenges in global AI governance, several strategies could be adopted. First is to establish a set of mechanisms for multilateral cooperation on the matter. For instance, the global community should consider building a multilateral institute dedicated to AI research to foster collaboration across countries and disciplines. Equally important is the creation of a unified ethical and legal framework, one that ensures that AI development aligns with human values, safeguards data privacy, and reduces social biases. It is also important to improve the transparency of our AI systems, which can facilitate a better understanding of AI decisionmaking and reduce our distrust of these models. Furthermore, we should strengthen regulations to ensure AI remains safe and reliable, and to prevent it from spinning out of human control.

 

Another focus should be on harnessing AI to promote social equity and serve the common good.

 

Last but not least, more efforts should be invested in training skilled AI professionals and advancing a deep fusion of AI with traditional industries, which can help us minimize the potential disruption AI may bring to the labor market.

 

These strategies can help us better address the many challenges posed by AI development while fully harnessing its potential to promote global fairness and sustainable development.

 

Future Trends in AI Development
From a technological standpoint, future AI systems will be capable of simultaneously processing a range of data types—including text, image, audio, and video—thereby offering a more diverse set of multimodal AI solutions.

 

With the continuous expansion of the Internet of Things (IoT), which describes a network of interrelated devices that connect and exchange data with other IoT devices and the cloud, the deployment of AI models will increasingly shift to local edge devices to enable real-time data processing and decisionmaking.

 

Breakthroughs in quantum computing are expected to significantly boost both the speed and quality of AI model training. Meanwhile, companies will move away from a reliance on singular LLMs toward integrated systems that combine diverse human roles, a variety of AI models, and different environmental settings—helping AI tackle more complex issues.

 

In terms of market growth, the global AI industry is experiencing exponential expansion, with a market size projected to grow at an annual rate of over 30 percent and reach $1.5 trillion by 2030. By 2035, China is projected to overtake the United States and become the world’s largest market for AI applications.

 

As for practical applications, AI is assuming a larger role in healthcare services, as its accuracy in providing medical diagnoses, imaging analysis, and personalized treatments continues to improve. It is also profoundly reshaping our educational systems, revolutionizing both how we learn and what we learn. In manufacturing, advances in digital twin technology—which works by digitally replicating a physical asset in the virtual environment—are expected to significantly reduce maintenance costs while boosting production efficiency. A further combination of AI and satellite remote sensing will enable better pest management and more precise control over agricultural activities.

 

On the policy front, governments worldwide are increasingly committed to supporting the growth of homegrown AI industries. China, for instance, has listed AI as one of the strategic emerging industries in its 14th Five-Year Plan, and has also established funds specifically dedicated to AI development. Meanwhile, global efforts to implement AI regulation and establish ethical frameworks are accelerating, greatly driving the standardization of AI governance.

 

Yet pitfalls abound. As AI grows ever more reliant on data, the protection of data privacy through privacy computing—which focuses on sharing the value of data without revealing the underlying data itself—will come to the forefront.

 

Furthermore, the widespread deployment of AI will require a more robust regulatory framework to ensure its sustainable growth.

 

In summary, AI is still in the early stages of becoming fully integrated into our daily lives, with technological expertise and industrial ecosystems only just beginning to take shape around the sector. Future breakthroughs in a set of revolutionary AI technologies, such as the aforementioned human-machine-intelligence system, will further unlock new growth potential for AI development.

 

In fact, the growth of the AI sector is unlikely to follow a linear path. Instead, advancements will unfold across multiple fronts, ranging from technological innovations and market expansion to the evolution of AI applications, policy support, and regulatory frameworks.

 

However, for all the transformative potential AI holds, it is crucial that we also focus on addressing the significant challenges that accompany its rise. This includes not only concerns surrounding data privacy and security but also the ethical dilemmas and societal implications that have emerged alongside the rise of AI.

Back to Table of Contents