By 2025, China's artificial intelligence industry will be full of vitality and highlights, with over 6000 AI companies and an expected core industry scale exceeding 1.2 trillion yuan. Against the backdrop of a new round of technological revolution and industrial transformation, the rapid development of artificial intelligence has not only given birth to new products and formats, but also led a systematic and fundamental change in research paradigms from the underlying logic, accelerating the transformation of research and development activities from traditional labor-intensive, trial and error, and linear progression to a new model of data-driven, intelligent simulation, collaborative parallel, and continuous iteration. Throughout the world, the application of artificial intelligence in research and development has moved from local tool applications to a new stage of full process and deep embedding of all elements. On the one hand, artificial intelligence is becoming an accelerator for cutting-edge scientific discoveries. In the fields of biomedicine, new materials, and new energy, artificial intelligence can compress the exploration process that originally took years into months or even weeks by efficiently processing massive amounts of scientific data and literature, significantly reducing the time and cost of basic research. On the other hand, artificial intelligence drives industrial technology development to achieve intelligent leaps. In the product design process, multiple design schemes can be automatically generated based on performance requirements; In the simulation testing phase, it can significantly reduce the number of physical prototype manufacturing and testing cycles; In the process optimization stage, the ability to analyze production data in real-time and independently search for the optimal parameter combination. This "data-driven+intelligent decision-making" model is driving research and development from experience dependence to intelligent driving. China has established a solid foundation in artificial intelligence technology and industrial applications, providing unique opportunities to lead the paradigm shift in scientific research. Firstly, China has a super large market and application scenarios, providing rich testing grounds and continuous feedback data for technology research and development. Secondly, in some fields, a relatively complete industrial chain has been formed, from computing power facilities, algorithm frameworks to industry applications, with an ecological foundation for collaborative innovation. Once again, the national strategy is accelerating its layout, the policy environment is continuously optimizing, and promoting the deep integration of artificial intelligence and the real economy has become a clear direction. However, it should also be recognized that truly integrating artificial intelligence into industrial innovation and research and development still faces a series of challenges. For example, there are still shortcomings in the basic layers such as key core algorithms, frameworks, and high-end chips, the construction of high-quality and standardized industry datasets lags behind, there is a shortage of composite talents who combine artificial intelligence technology and industry knowledge, and the R&D management mechanism and culture that adapt to agile innovation and fault-tolerant trial and error have not yet been widely formed. Systematically promoting the deep empowerment of artificial intelligence in industrial innovation requires collaborative efforts from multiple dimensions such as technological breakthroughs, ecological construction, factor support, and institutional guarantees. Strengthen basic research and development, and build a solid foundation for intelligent research and development. The fundamental transformation of R&D paradigm cannot be achieved without a solid technological foundation. We need to continue to increase investment in original innovation of underlying algorithms such as machine learning, knowledge graphs, and large models, and encourage research institutions and enterprises to jointly build high-level open source frameworks and platforms for artificial intelligence. We will focus on breaking through hardware bottlenecks such as high-end artificial intelligence chips and advanced computing architectures, and enhance our ability to independently and controllably supply intelligent computing power. Build a collaborative ecosystem and connect the R&D innovation chain. Encourage "chain owner" enterprises, universities, new research and development institutions, and startups to form innovation consortia, and carry out in-depth cooperation around common technology research and development, pilot testing, and scenario applications. Promote the construction of national level innovation centers or open platforms that integrate artificial intelligence, provide model services, computing power support, and technical consulting, and lower the application threshold for small and medium-sized enterprises. Cultivate versatile talents and innovate organizational management mechanisms. Talent and organization are the key guarantees for the implementation of paradigm shift. We need to accelerate the improvement of the discipline construction in the field of artificial intelligence, and promote the joint training of "AI+X" composite talents who not only understand artificial intelligence technology but also master relevant industry knowledge between universities and enterprises. Encourage the establishment of independent artificial intelligence research and development innovation units or laboratories within enterprises, giving them greater technological decision-making power and resource scheduling flexibility. Transform the traditional linear project management model, promote agile R&D and flat management mechanisms that adapt to rapid trial and error and continuous integration, and create an innovative culture that encourages exploration and tolerates failure. Expand the application scenarios and improve the policy support system. The broad application scenarios are an important driving force for the maturity of driving technology and the deepening of paradigms. We should focus on key areas such as manufacturing R&D and design, biopharmaceutical discovery, and new material synthesis, and systematically release a list of key scenarios for AI enabled R&D, attracting various forces to tackle them through methods such as "unveiling the list and leading the way". Improve the first purchase, first use, and insurance compensation mechanism for innovative products, providing early support for the market-oriented application of AI driven research and development achievements. At the policy level, it is necessary to study and formulate statistical evaluation, fund management, ethical safety, and intellectual property protection rules that are adapted to the characteristics of artificial intelligence research and development, in order to create a stable, inclusive, and sustainable institutional environment for the paradigm shift in research and development. (New Press) (Authors Zhang Linshan and Gong Piming, respectively, are researchers and associate researchers at the Macroeconomic Research Institute of the National Development and Reform Commission)