I. Guiding Principles

This plan aims to accelerate the implementation of the innovation-driven development strategy and move faster towards self-reliance in advanced science and technology. It adheres to the principle of encouraging independent innovation, staying problem-oriented, making coordinated plans, and advocating open approach for innovation. By doing all these, Beijing will hopefully leverage its advantageous resources for AI innovation to further enhance its global influence and push the development of its AI industry to the forefront of the world.

II. Objectives

Beijing’s AI industry and relevant technological innovation efforts will have moved into a new development stage by 2025 and an AI innovation hub with global influence will have basically taken shape, according to the plan. 

Focus will be laid on a number of cutting-edge sectors so as to propel development through technological innovation. 

Breakthroughs will be made in a series of projects to deliver domestic products that can be alternatives to imported counterparts and in some key technological fields.

Some industrial clusters will be established to realize upgrading of the AI industry. 

A number of model projects will be recognized and promoted and new applications will be formed based on specific scenarios. 

The city will put in place first-class innovation environment and make new achievements in the development of an industrial ecosystem. 

III. Main Tasks

(I) Work for breakthroughs in key technologies

1. Pursue breakthroughs with innovation in cutting-edge fields of fundamental theoretical research. Fundamental theoretical frameworks will be developed for new-generation AI. Support will be given continuously to innovators such as new R&D centers to work in general intelligent agents, scientific computing and other cutting-edge fields.

2. Take the lead in pursuing innovations in key and core technologies. Beijing will seek breakthroughs in fundamental platform technologies, such as distributed deep learning frameworks, develop key algorithms for federated deep learning spanning data and knowledge. and encourage open-source initiatives and sharing.

3. Enhance innovation in trustworthy AI. Work will be done on system robustness, explainability, fairness, and security to develop strategies and frameworks to bolster privacy protection regarding sensitive data and establish a categorized and tiered system trustworthiness evaluation and test system.

(II) Go all out to lay a solid foundation

4. Seek breakthroughs in developing domestic AI chips. Work will be done to develop high computing power training and low-power edge AI chips, explore new architecture routes, and increase domestic computing power supply.

5. Work harder on R&D of deep learning frameworks. R&D efforts will be made on efficient distributed computing to develop tools that encompass the whole process from multi-type model development to inference, and push for compatibility between domestically made fundamental software and hardware.

6. Improve coordination in computing power distribution. Beijing will promote collaboration in computing power among different regions by building or renovating a group of commercial public computing centers and launching the computing power partnership program.

7. More work will be done for public data sharing. The public data stock will be processed more effectively, public data platforms of various kinds will be improved and explorations will be made towards authorized operation of public data platforms and build open high-quality training datasets.

(III) Accelerate the construction of industrial clusters

8. Build technology-based industry ecosystems. Open-source and open-access platforms for large model algorithms and tools, and domestic full-stack AI innovation consortiums will be established to attain an independent and manageable technology-based industrial ecosystem.

9. More efforts will be made to build a well-structured echelon of enterprises. Beijing will facilitate the in-depth integration of industry, education institutions and research institutes; enhance the cultivation of unicorns, establish benchmark incubators, and attract competitive teams to Beijing.

10. Improve multi-dimensional services for enterprises. Beijing will enhance its ability to provide better services to enterprises, put in place service mechanisms, and broaden the range of services available.

(IV) Accelerate the development of application scenarios

11. Explore scenario-empowered opening up. A number of benchmark application scenarios will be created to foster the iteration of new technologies and the advancement of innovative applications, and promote more in-depth integration between the innovation chain, the industry chain, the capital chain and the talent chain.

12. Support the building of smart cities. Support will be given to Haidian District for Urban Brain 2.0 and efforts will be made to promote projects like Beijing Municipal Advanced Autonomous Driving Demonstration Zone 3.0.

(V) Develop with continuous efforts an ecosystem that nurtures innovation

13. Promote the building of a talent hub. Policies and measures will be tailored to attracting AI talent, encourage enterprises to attract and train technology talent of various levels, and support higher education institutions in exploring interdisciplinary talent training.

14. Create an environment conducive to innovations in the field of AI. AI industry clusters will be built and platforms for international exchange and cooperation established. Support will be provided for early-stage investment in key and core technologies and for enterprises in getting listed.

15. Explore inclusive and prudential supervision. Innovation in supervisory policies and procedures will be promoted, and mechanisms for regular communication, services, and guidance established. Efforts will be made to facilitate security assessment and algorithm filing, and raise the security awareness of innovators.

16. Enhance governance related to technology ethics in the AI industry. Research will be strengthened in regard to ethical security regulation and relevant social governance practices, and technology ethics examination and the relevant training sessions will be launched. An expert review mechanism will be established to further check the ethical examination results of high-risk sci-tech activities.

IV. Support Measures

1. Reinforce organizational leadership. Beijing will establish a special task force to enhance coordination, improve resource allocation, and bolster the innovation in the development of the AI industry in the city.

2. Promote institutional innovation. Institutional reforms will be promoted, competition will be launched for the completion of strategic tasks, explorations will be made to form a market-oriented mechanism for the commercialization of technological achievements, and  efforts will be made to accelerate the building of a collaborative innovation system encompassing the AI industry, education institutions, and research institutes.

3. Improve support policies. Beijing will strengthen innovation in policies and provide greater financial support, implement dedicated sci-tech development plans for the field of AI, improve the investment mechanism for industrial advancements, and put in more resources.

4. Implement all related policies in well-coordinated manner. Annual work plans will be formulated to help strengthen support mechanisms and real-time information monitoring, and proactively conduct international exchanges and cooperation.