AI technological innovation ushers in a new period of opportunities

Economic Reference News Reporter Li Zhiyong

This year’s government work report proposes deepening and expanding “AI+” initiatives, promoting the rapid adoption of next-generation intelligent terminals and intelligent agents, and encouraging large-scale commercial application of AI in key industries. It also aims to cultivate new business models and formats rooted in intelligence. Experts say that with continuous upgrades and implementation of domestic AI technology, China’s AI innovation is entering a new phase of opportunity.

Policy Support Accelerates Large-Scale AI Deployment

Li Lecheng, Minister of Industry and Information Technology, said in an interview during the Two Sessions that efforts should be made to promote the development of AI computers, AI smartphones, and smart home devices to better meet people’s needs for a better life. He expressed confidence that more world-class intelligent products will be produced on this fertile land.

In recent years, policies have consistently strengthened support for AI technology. The “Opinions of the State Council on Deepening the Implementation of ‘AI+’ Action” released in August 2025 explicitly states that by 2027, AI will be broadly and deeply integrated into six key areas, with the adoption rate of next-generation intelligent terminals and intelligent agents exceeding 70%. By 2030, the penetration rate will surpass 90%, making the intelligent economy a major growth engine for China’s economic development, promoting technological inclusiveness and sharing of achievements.

This set of quantitative goals has shifted AI development from isolated breakthroughs to large-scale implementation across the entire industry chain.

In December 2025, the Central Economic Work Conference proposed to deepen and expand “AI+” initiatives and improve AI governance. In February 2026, during the 18th special study session of the State Council, it was emphasized that understanding and grasping the development trend of AI is crucial, and efforts should be made to break through the entire AI chain and implement AI across all scenarios, unlocking greater development potential.

Yin Ximing, Director of the Public Management Department at Beijing Institute of Technology and Deputy Director of the Guangdong-Hong Kong-Macao Greater Bay Area Digital Economy and Technology Innovation Research Institute, said that China has produced influential large models and applications such as DeepSeek, Doubao, and Qianwen, and has made forward-looking deployments and market explorations in intelligent terminals and intelligent agents like smartphones and computers. During the 14th Five-Year Plan period, AI technology will be deeply integrated into core business operations of enterprises and industries, becoming a driving force for organizational intelligence evolution, disruptive technological innovation, and global industrial paradigm shifts. To further promote innovative applications of intelligent terminals and agents, the focus should be on innovation-led, scenario-driven development, with enterprises as the main actors, supported by inclusive and cautious governance. Accelerating the large-scale application of new technologies, products, and scenarios such as intelligent agents, edge-side AI, and smartphone assistants is essential.

Rapid Evolution of New Tech Routes and Terminal Ecosystems

Continuous policy support has fostered vibrant industry growth. In January 2025, DeepSeek released its reasoning model DeepSeek-R1 as open source. Its performance, characterized by low cost and near-top-tier closed-source models, sparked international academic and industry discussions, even prompting reevaluation of computing power and model routes in overseas markets. By the end of 2025, ByteDance unveiled a preview of Doubao smartphone assistant, combining leading model capabilities with hardware to bring technological innovation into daily scenarios.

Recently, Google Gemini and Samsung collaborated to launch a new AI smartphone. Through integrated development of large models, hardware, and operating systems, they promote cross-application task execution and scenario-based intelligent services, forming a “model-terminal-ecosystem” paradigm. Giants like OpenAI and Microsoft are also actively exploring cutting-edge innovations such as GUI-based intelligent agents. Experts say this indicates the future direction of global intelligent terminals. The next-generation mobile ecosystem competition will no longer be about hardware or software alone but about system-level intelligence, full-chain services, and inclusive applications.

Liu Dian, Associate Researcher at Fudan University’s China Research Institute, said that global AI competition has shifted from technological R&D and product deployment to a deeper stage involving ecosystem construction and rule-making. China has developed unique advantages in large models, terminal manufacturing, and application scenarios. It should encourage independent, controllable new technologies and products to accelerate scenario-based deployment, driven by application and validated by market value. Domestic mobile intelligent agents like Doubao, which leverage large models for understanding, reasoning, and execution, are providing automated services for complex tasks in high-frequency public needs—an example of AI moving from labs to everyday life. Deep cooperation between hardware manufacturers and large model companies, following open, collaborative, secure, and scenario-prioritized industry paths, can resonate with international trends while building a digital ecosystem tailored to local needs.

Secure Foundations Drive Sustainable AI Innovation

The rise of intelligent agents is reshaping human-computer interaction, application distribution, and business profit models. While bringing convenience, data security concerns are also prominent. Experts say that AI agents can technically ensure user privacy and data security, and a security foundation built through policies, technology, and regulations will support healthy AI development.

Sha Lei, Professor at Beijing University of Aeronautics and Astronautics Artificial Intelligence Research Institute, stated that AI agents are now capable of protecting user privacy and data security at the technical level. Industry consensus is forming around security design: all operations are based on user authorization; task decomposition and execution are transparent, perceptible, and controllable; data transmission and storage are strictly encrypted, with personal information anonymized and desensitized; and privacy impact assessments and compliance checks are conducted by specialized internal teams to reduce risks of data leaks.

陶乾, Director of the Intellectual Property Innovation and Competition Research Center at China University of Political Science and Law, believes that AI agents operate based on system-level permissions authorized by users, with behavior limited to the scope of user interface interactions. Their actions should be regarded as “a technical realization of user intent,” not independent third-party actions.

Professor Zhang Jiyu of Renmin University Law School said that in cases where privacy and data protection measures are reasonable and do not improperly harm third-party platform interests (such as ride-hailing assistance), platform approval is not required. However, in high-risk areas or those significantly affecting platform rights (such as game cheats or bot armies), cautious regulation is necessary.

Fang Xingdong, Executive Dean of the School of Media and International Culture at Zhejiang University, emphasized that we should not cling to existing patterns or use a “one-size-fits-all” approach to stifle innovation. Regulators should create an inclusive and cautious environment for AI new formats, encouraging internet giants to shift from closed “data land grabbing” to open technology and service ecosystems, jointly embracing the next stage of the digital era amid turbulence and testing.