This year's government work report proposes to deepen the expansion of "artificial intelligence+", promote the accelerated promotion of new generation intelligent terminals and agents, promote the commercial and large-scale application of artificial intelligence in key industries, and cultivate intelligent native new business models. Experts say that with the continuous upgrading and implementation of domestic AI technology, China's AI technology innovation is ushering in a new round of opportunities. Li Lecheng, Minister of Industry and Information Technology, stated in an interview at the Two Sessions that efforts should be made to promote the development of AI computers, AI smartphones, and smart homes, in order to better meet the needs of the people for a better life. He said that we have confidence and firmly believe that more world-class intelligent products will be produced on this hot land. In recent years, the policy level has continuously strengthened support for AI technology. The Opinion of the State Council on Deepening the Implementation of the "Artificial Intelligence+" Action, released in August 2025, clearly states that by 2027, we will take the lead in achieving extensive and deep integration of artificial intelligence with six key areas, and the penetration rate of new generation intelligent terminals, intelligent agents and other applications will exceed 70%. By 2030, the penetration rate of new generation intelligent terminals, intelligent agents and other applications will exceed 90%, and the intelligent economy will become an important growth pole for China's economic development, promoting technology inclusiveness and achievement sharing. The setting of this quantitative goal has driven AI technology to shift from "single point breakthroughs" to "large-scale implementation of the entire industry chain". In December 2025, the Central Economic Work Conference proposed to deepen and expand "artificial intelligence+" and improve artificial intelligence governance. In February 2026, during the 18th special study session of the State Council, it was also pointed out that it is necessary to deeply understand and grasp the development trend of artificial intelligence, promote breakthroughs in the entire chain and full scenario implementation of artificial intelligence, and unleash greater development potential. Yin Ximing, Director of the Department of Public Administration at Beijing Institute of Technology and Vice Dean of the Guangdong Hong Kong Macao Greater Bay Area Digital Economy and Technology Collaborative Innovation Research Institute, stated that China has emerged with globally influential large-scale model products and applications such as DeepSeek, Doubao, and Qianwen, and has also carried out forward-looking deployment and market-oriented exploration in intelligent terminals and agent applications such as smartphones and smart computers. During the 15th Five Year Plan period, artificial intelligence technology will be deeply integrated into the core business of enterprises and industries, becoming the driving force for organizational intelligence evolution, technological disruptive innovation, and global industrial paradigm change. To further promote innovative applications of intelligent terminals and agents, we must adhere to innovation led and scenario driven approaches, with enterprises as the main body and inclusive and prudent governance as the guarantee, and accelerate the large-scale application of new technologies, products, and scenarios, including intelligent agents, end side AI, and smartphone assistants. The continuous strengthening of new technological routes and policies for the rapid evolution of terminal ecosystems has spurred the vigorous vitality of the industry. In January 2025, DeepSeek released the inference model DeepSeek-R1 and made it open source. Its "low-cost+close to top closed source model" performance sparked international academic and industrial discussions, and even led to a repricing of computing power and model routes in overseas markets. At the end of 2025, ByteDance will release the technical preview version of Doubao mobile phone assistant, and try to put technical innovation into daily scenarios through the combination of leading model capabilities and hardware terminals. Not long ago, Google Gemini and Samsung collaborated deeply to launch a new AI smartphone. By integrating large models with hardware terminals and operating systems, both parties promote cross application task execution and scenario based intelligent service implementation, forming a paradigm of "model terminal ecology". OpenAI、 Microsoft and other giants are also actively exploring global innovation frontiers such as GUI intelligent agents. Experts say that this represents the evolution direction of global intelligent terminals. The competition in the next generation of mobile ecology is no longer a competition of single hardware or software, but a comprehensive competition of system level intelligence, full link services, and inclusive applications. Liu Dian, Associate Researcher at the China Research Institute of Fudan University, stated that the current global AI competition has entered a deep stage from technology research and development, product landing, to the construction of industrial ecology and the competition for rule making rights. China has formed unique advantages in areas such as AI big models, terminal manufacturing, and application scenarios. It should encourage the acceleration of scenario based implementation of independently controllable new technologies and products, driving innovation through applications and verifying value through the market. Local mobile intelligent agents, represented by Doubao mobile assistants, rely on large models to achieve integrated understanding, reasoning, and execution, providing complex task automation services for high-frequency needs of people's livelihoods. This is a typical practice of AI technology moving from the laboratory to mass life. Domestic hardware manufacturers and large model enterprises engage in deep cooperation, taking an open and collaborative, secure and controllable, and scenario first industrial path, which can resonate with international cutting-edge trends and build a digital ecosystem that meets local needs. The rise of intelligent agents is reshaping the way human-machine interaction, application distribution logic, and business benefit distribution pattern, as the security foundation drives the sustainable development of AI innovation. While intelligent agents bring more convenience to people, issues such as data security have also become a focus of attention. Experts suggest that intelligent agents can ensure user privacy and data security in terms of technology, and the security foundation constructed by policies, technology, regulations, etc. will also promote the sustained and healthy development of AI technology. Professor Sha Lei from the Institute of Artificial Intelligence at Beihang University stated that intelligent agents have the ability to protect user privacy and data security at the technical level. The technological design of intelligent agents in terms of security is forming an industry consensus: all operations are based on user authorization, and task disassembly and execution are visible, perceptible, and manageable throughout the process; Strictly encrypt the transmission and storage of user data, and anonymize and anonymize personal information; By conducting multiple protections such as privacy impact assessment and compliance testing through an internal professional team, the risk of privacy leakage can be reduced from a mechanistic perspective. Tao Qian, director of the Intellectual Property Innovation and Competition Research Center at China University of Political Science and Law, believes that the operation of AI agents relies on system level permissions authorized by users in advance, and their operational boundaries are limited by the interface interaction scope authorized by users. Their behavior should be recognized as a "technical implementation mechanism of user will" rather than an independent third-party subject behavior. In the view of Professor Zhang Jiyu from the Law School of Renmin University of China, for situations where reasonable privacy and data protection measures have been taken and the interests of third-party platforms will not be improperly harmed (such as helping users hail a taxi), platform consent is not required. For areas with high social risks or significant impact on platform rights (such as game cheating and robot navy), careful control is necessary. Fang Xingdong, Executive Vice Dean of the School of Media and International Culture at Zhejiang University, stated that innovation attempts cannot be completely stifled by sticking to the existing pattern and using a "one size fits all" approach. The regulatory level should create an inclusive and prudent competitive environment for AI's new business, force Internet giants to move from a closed "data enclave" to an open technology and service ecosystem, and jointly welcome the next step of the digital era in shocks and stress testing. (New Society)