With the rapid advancement of new-generation AI technologies, intelligent agents are accelerating their in-depth integration with cyberspace and the physical world, transforming people's production methods and daily life as well as social governance models.

To promote innovative development of AI agents and advance "AI Plus" action, the Cyberspace Administration of China (CAC), the National Development and Reform Commission and the Ministry of Industry and Information Technology have jointly released an implementation guideline.

The guideline defines AI agents as intelligent systems equipped with autonomous perception, memory, decision-making, interaction and execution capabilities, serving as an important form of AI products and services.

It puts forward measures in four key aspects:

•Consolidating the foundation for development, improving the technical infrastructure, and establishing standards and protocols;

•Upholding the bottom line of security, clarifying product norms, guarding against potential safety risks, improving the governance system, and strengthening industrial self-regulation;

•Strengthening application-driven development. Centered on scientific research, industrial development, consumption promotion, people's well-being, and social governance, it sets forth 19 typical application scenarios;

•Building an innovation ecosystem, boosting industrial cooperation, and scaling up application promotion.

The CAC noted AI agent products such as mobile assistants, terminal smart stewards, and cloud-based AI agents have emerged at a rapid pace in recent years and entered large-scale application, greatly facilitating people's work and daily lives.

However, the high autonomy and high permission level of intelligent agents also bring security risks including privacy leakage, unauthorized operation, and behavioral malfunction. Therefore, it is imperative to balance development and security to advance the standardized application and innovative development of intelligent agents.

The guideline proposes a people-centered approach, AI for good, multi-party collaborative governance, and sound and prudent development. This will foster an institutional environment that ensures standardized development while encouraging innovation, thereby promoting healthy and orderly growth of intelligent agents.

Clear product guidelines should be established. This means improving policies, regulations and ethical norms, as well as managing permissions and regulating the behaviors of intelligent agents to guide the R&D of related products.

At the same time, the governance system needs to be improved by building a classified and tiered framework, and professional services such as risk monitoring, early warning, testing, and certification for AI agents strengthened.

An evaluation indicator system will be established for the development of intelligent agents. Monitoring, assessment, rolling implementation and dynamic adjustment will be strengthened for the standardized application and innovative development of AI agents.