China's MIIT Unveils AI+Software Action Plan: 2028 Targets for Intelligent Programming & Agent Software

Overall Requirements

Guided by Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era, the plan adheres to principles of application-driven, innovation-driven, security-controllable, and ecological collaboration. It aims to reshape software and IT services with AI, drive industrial efficiency through intelligent programming, coordinate software product intelligent upgrades and intelligent agent cultivation, enhance IT service quality, solidify intelligent foundations, and foster a favorable environment for a new development pattern of efficient production, intelligent products, inclusive services, and sound ecology.

2028 Targets

Significant improvement in intelligent level of software and IT services industry.

Cultivate a batch of high-level intelligent programming tools and intelligent development platforms.

Promote application covering 20,000+ scale-up software enterprises.

Achieve significant results in key software intelligent upgrades.

Implement 100 software enterprise intelligent transformation projects cumulatively.

Accelerate emergence of new software forms represented by intelligent agent software.

Build 100 intelligent agent software benchmark applications in key industries.

Upgrade IT services toward intelligent and value-added, forming industry deep-application solutions and typical scenarios.

Make positive progress in open collaborative open-source community construction, incubating 5+ quality open-source projects.

2030 Vision

Achieve a new leap in integrated development of AI with software and IT services; fully realize intelligent upgrade of key software; make new business forms such as intelligent programming, intelligent agent software, and intelligent services become new growth poles; build several open-source communities and industrial clusters with international influence; propel China's software and IT services industry to seize the high end of the global value chain.

Key Tasks

1. Advance Software Production Reform, Accelerate Toward "Intelligent-Driven"

(1) Optimize Intelligent Programming Toolchain : Develop intelligent-agent-driven intelligent programming tools, deeply optimized for software development scenarios, enhancing project-level, full-process autonomous development capabilities. Promote deep adaptation and integration of intelligent programming tools with code hosting platforms, cloud services, and open-source communities to perfect the software development and deployment ecosystem. Support qualified enterprises in large-scale adoption to continuously improve model and tool maturity. Develop user-friendly intelligent programming tools for the masses to lower usage thresholds.

(2) Accelerate Software Enterprise Intelligent Transformation : Support qualified localities in organizing software enterprise intelligent transformation, guiding enterprises to increase investment in intelligent programming applications and transform software development processes. Encourage localities to coordinate existing funds and policy channels, using code quality and R&D efficiency as key criteria, prioritizing support for enterprises with outstanding results. Select regions and enterprises with good foundations for pilot trials, summarize replicable and promotable experiences, and extend intelligent transformation to software development links in manufacturing, government affairs, finance, medical, education, transportation, and other key industries.

(3) Empower Software Security with Intelligent Programming : Support enterprises in using large models and intelligent programming tools to conduct vulnerability mining, automatic defect identification, and risk grading assessment for existing software, improving discovery and remediation capabilities for high-risk security hazards. Promote embedding intelligent security detection throughout the software development lifecycle for real-time hazard discovery and timely handling. Strengthen intelligent security testing means such as penetration testing and attack-defense simulation to enhance software security verification capabilities.

2. Accelerate Software Product Intelligent Upgrade, Vigorously Promote "Intelligent Partners"

(4) Enhance Basic Software Intelligent Level : Targeting operating systems, databases, and other basic software, support qualified enterprises in implementing intelligent upgrades around intelligent agent scheduling, performance tuning, operation and maintenance management, and security protection. Support basic software enterprises in collaborating with chip and server hardware vendors for underlying adaptation and software-hardware collaborative optimization joint research, building 15+ adaptation centers to comprehensively improve industrial verification capabilities.

(5) Drive Industrial Software Intelligent Transformation : Promote deep integration of general industrial software such as CAD, CAE, and EDA with AI, implementing intelligent upgrades for drawing, design, and simulation scenarios, exploring next-generation industrial software development paths. Support qualified industrial enterprises in partnering with software enterprises to encapsulate industrial experience and process parameters into reusable model libraries and industrial components, developing industry-specific intelligent industrial software. Advance integration of large model capabilities into industrial internet platforms to improve data analysis, production scheduling, and multi-agent collaborative decision-making levels.

(6) Accelerate Application of New Intelligent Software Products : Enrich software intelligent functions across production and life domains, continuously expanding coverage of software intelligent upgrades. Support software enterprises in joint innovation with terminal manufacturers such as smartphones, smart homes, and intelligent connected vehicles to accelerate lightweight model and terminal software fusion deployment applications, and explore embodied intelligent system software development. Cultivate quantum computing, brain-like computing, and other system software, exploring future software directions represented by intelligence-native.

3. Cultivate Intelligent Agent Software New Business Forms, Accelerate "Continuous Evolution"

(7) Solidify Intelligent Agent Software Technical Foundation : Strengthen engineering technology research on intelligent agent operation frameworks, improving reliable execution of complex tasks and multi-agent collaborative operation capabilities. Support intelligent agent software development platform construction, perfecting toolchains for R&D, testing, deployment, and operation. Enhance integration and adaptation levels with existing software systems to promote large-scale application of intelligent agent software.

(8) Cultivate Intelligent Agent Software Products : Develop cross-scenario applicable general intelligent agent software for information processing, office collaboration, life services, and other diversified scenarios. Promote intelligent agent empowerment of Internet applications and services, accelerating development of terminal intelligent agent software with lightweight reasoning and privacy protection capabilities. For high-value industry scenarios, integrate knowledge graphs and data resources to build vertical domain professional intelligent agent software. Focus on core industrial links to develop safe, reliable, and verifiable industrial intelligent agents, strengthen interaction adaptation with industrial control system software and enterprise core business systems, and deeply integrate into the whole production and operation process.

(9) Build Intelligent Agent Software Application Market : Promote construction of intelligent agent software app stores and skill package resource libraries, standardize platform listing review and operation management, and promote wide application of intelligent agents and skill packages. Guide various innovation entities to develop high-quality professional skill packages and knowledge bases, encourage software enterprises to develop functional components for intelligent agents, and precipitate industry knowledge. Cultivate intelligent agent open-source zones to build a multi-party collaborative, skill-sharing open ecology.

4. Expand Intelligent Software Services, Actively Transform "Service Models"

(10) Innovate Intelligent Software Service Models : Support software enterprises in building end-to-end intelligent application service chains based on large models and intelligent agent technologies. Encourage provision of ubiquitous, accessible, and easy-to-use intelligent software services for consumer markets, forming a virtuous cycle of "scale application — technology iteration — service inclusiveness." For high-value vertical scenarios, develop specialized intelligent software services, support procurement of large model and intelligent agent services, and improve industry model and intelligence usage quality and efficiency.

(11) Cultivate and Strengthen Intelligent Software Service Entities : Guide software enterprises to transform into AI application service providers, enhancing continuous service and value delivery capabilities. Implement AI SME entrepreneurship support plans, focusing on supporting lightweight, specialized new software service entities, cultivating a batch of startup enterprises and specialized "little giant" enterprises. Encourage qualified localities to cultivate "AI+Software" innovation and entrepreneurship agglomeration zones, providing coordinated venue, computing power, and other supporting development resources.

(12) Accelerate Intelligent Software Service Landing Applications : Encourage software enterprises to pair with key industry users, targeting common typical scenarios to land first-use, real-use, replicable and promotable solutions. Hold AI-empowered software "deep dive" activities, advancing demand docking for high-value scenarios in manufacturing, finance, and other industries by sector. Rely on various industry public service platforms to provide lightweight, low-cost solutions for SMEs.

5. Solidify Software Intelligent Development Foundation

(13) Promote Software Open-Source Collaborative Innovation : Relying on national open-source foundations and AI open-source communities, guide enterprises to open achievements in intelligent programming, software product intelligence, intelligent agents, etc. Support key open-source projects for community first release, establish open-source zones for intelligent operating systems and other fields, strengthen autonomous open-source license application and compliance guidance. Establish quality open-source project selection and cultivation mechanisms, encourage exploration of diversified business models, and promote deep application of quality projects in manufacturing, finance, and other fields. Improve open-source contribution incentives, support universities, research institutes, and enterprises in incorporating open-source contributions into scientific research evaluation and talent recognition systems, and explore using open-source contributions as important reference indicators for discovering quality software SMEs.

(14) Strengthen Computing Power Service Support for Software Enterprises : Improve intelligent computing cloud service systems, develop integrated services of computing power, models, and software tools, and promote intensive resource utilization. For software enterprise computing needs, support development of elastic, flexible, on-demand scalable computing power services. Make good use of inclusive support policies such as computing power vouchers to reduce software enterprise computing costs, and strengthen computing power supply-demand matching and cloud usage guidance for SMEs.

(15) Build Software Industry High-Quality Datasets : Focusing on key scenarios such as code generation, intelligent testing, and intelligent operation and maintenance, support construction of software industry high-quality datasets to promote mutual enhancement of data and models. Improve toolchains for data annotation, cleaning, and synthesis, guiding enterprises to enhance software engineering data governance capabilities. Rely on open-source communities and code hosting platforms to strengthen data sharing and utilization.

6. Optimize Software Industry Development Environment

(16) Advance Key Standard Formulation in Software Field : Around intelligent programming, software product intelligence, intelligent agents, and intelligent services, coordinate advancement of key standard formulation. Formulate intelligent programming capability maturity grading evaluation standards, develop intelligent agent interface standards for software-hardware collaboration. Research and formulate intelligent service standards such as "Model as a Service" and "Intelligent Agent as a Service," accelerating application in manufacturing and other key industries.

(17) Strengthen Software Talent Cultivation and Supply : Facing new occupation and post demands in software field, strengthen composite talent cultivation. Support universities in updating relevant disciplines and majors in line with industrial development, deepening interdisciplinary integration. Encourage school-enterprise cooperation in offering industry-education integration courses and co-building talent training bases, implement excellent engineer cultivation plans and vocational education high-skilled talent cluster cultivation plans. Build international exchange platforms to strengthen introduction and cultivation of overseas top talents.

(18) Promote Employment-Friendly Development of Software Industry : Adhere to employment-friendly orientation, guide software enterprises to optimize human-machine collaborative production organization modes while improving intelligent levels, simultaneously implement post transformation and skill training, and actively develop new posts. Expand new employment space through development of new business forms such as intelligent agent software and intelligent services. Strengthen new skill training related to software intelligence, promote software practitioners' transformation to high-value-added posts. Strengthen assessment of AI impact on software field employment, improve risk response mechanisms.

(19) Enhance Software Field Security Assurance Capabilities : Strengthen security review of AI-generated code, prevent new attacks such as malicious instruction injection in intelligent programming tools, and strengthen access control for development tools and code repositories. Research intelligent agent identity identification, trusted interconnection, data security, behavior control, and other security technologies, promote establishment of intelligent agent security management norms covering development, deployment, application, and other links. Facing AI integration trends, improve software product and service quality management and evaluation norms. Strengthen intellectual property compliance governance of AI-generated code.

7. Strengthen Organizational Guarantee

Strengthen organizational implementation, enhance cross-departmental and central-local coordination linkage, rely on "open bidding" and other mechanisms to promote key task landing. Increase policy support, coordinate various fund channels, first-version software product recognition, and other policies, giving support to qualified "AI+Software" projects. Use special fiscal funds to accelerate key technology R&D, infrastructure construction, and scenario applications. Co-build prosperous ecology, support holding AI innovation application competitions and other industry activities, expand multi-level international cooperation, and create an open and win-win industrial atmosphere.