20 Chinese Tech Firms Sign AI Agent Self-Discipline Pact to Promote Safer AI

Twenty Chinese technology organisations have joined forces to introduce a self-discipline pact aimed at promoting the responsible development and use of artificial intelligence agents.

The agreement, called the Agent Development Self-Regulation Convention, was officially released on November 4, 2026, during the World Internet Conference Wuzhen Summit.

The initiative brings together the China Academy of Information and Communications Technology (CAICT) and major technology companies, including Huawei, Alibaba, Tencent, Xiaomi, Baidu, Ant Group and ByteDance’s Douyin.

The agreement outlines eight commitments covering AI agent trust, infrastructure, interoperability, security, governance and user protection.

The announcement comes as AI agents are becoming more capable of completing tasks, interacting with digital services and making decisions with limited human involvement.

What Is the AI Agent Self-Discipline Pact?

Unlike traditional chatbots that mainly respond to user questions, AI agents can perform multiple steps to complete a task.

For example, an AI agent could research a topic, collect information, compare options, interact with software and prepare a final report. Some agents can also connect with other agents and external systems.

These capabilities create new opportunities for businesses and individuals. However, they also introduce risks involving privacy, security, unauthorised actions and accountability.

China’s new AI agent self-discipline pact aims to address these challenges through voluntary industry commitments.

The agreement encourages participating organisations to build trustworthy AI systems, improve cooperation between different platforms and ensure that AI agents operate within appropriate rules.

Eight Key Commitments in the Agreement

The convention outlines eight areas where participating organisations have agreed to support responsible AI agent development.

1. Build a Trusted AI Agent System

Companies are encouraged to establish reliable identity and trust mechanisms for AI agents. This includes improving identity verification, access permissions and controls over agent activities.

2. Develop Shared Infrastructure

The agreement encourages the development of infrastructure that allows AI agents to discover and communicate with one another across different platforms.

3. Improve Interoperability

Different AI systems often use different technologies and communication methods. The pact encourages better compatibility so agents from different providers can work together more effectively.

4. Strengthen Security Protection

Organisations are expected to improve security measures and protect AI agents against technical vulnerabilities, external attacks and unsafe behaviour.

5. Improve AI Governance

The agreement promotes governance mechanisms, compliance testing and built-in rules to ensure AI agents follow applicable laws and accepted standards.

It also calls on companies to prevent the use of AI agents for algorithmic exploitation and illegal or harmful online activities.

6. Protect User Rights

User privacy and control are important parts of the agreement. Organisations are encouraged to protect personal information and respect users’ rights when AI agents handle tasks and make decisions.

7. Expand AI Across Industries

The participating organisations aim to accelerate the use of AI agents in different industries. The focus is on practical applications that can improve business operations and productivity.

8. Make AI Technology More Accessible

The final commitment encourages open-source development, shared research, international standards and training opportunities.

The aim is to make AI agent technology easier for developers and organisations to access and deploy.

Major Technology Companies Are Involved

The agreement includes several of China’s largest technology organisations.

Some of the participating names include:

• Huawei: A major technology and telecommunications company.
• Alibaba: An internet and cloud computing company with a growing AI ecosystem.
• Tencent: A technology company known for its digital platforms and AI development.
• Xiaomi: A consumer electronics and smart device manufacturer.
• Baidu: A technology company developing AI models and services.
• Ant Group: A financial technology company.
• Douyin: The Chinese short-video platform operated by ByteDance.
• OPPO and Honor: Smartphone and consumer technology manufacturers.

The participation of companies from different technology sectors highlights the importance of shared AI standards. AI agents are expected to operate across devices, applications, cloud platforms and business systems rather than remain limited to individual products.

Why AI Agent Safety Is Becoming Important

AI agents are moving from experimental tools into real business environments.

Companies are exploring their use in customer service, software development, financial operations, logistics, healthcare and other areas.

However, an agent that can independently interact with systems may create risks if it has excessive permissions or receives incorrect instructions.

For example, an agent could expose private information, perform an unauthorised action or interact with a service in an unintended way.

There are also concerns about the misuse of agents for automated scams, online manipulation, fraudulent activities and other harmful operations.

The Chinese agreement specifically addresses these risks by encouraging stronger identity verification, security monitoring, permission controls and user protection.

It also highlights the need to prevent AI agents from being used for algorithmic exploitation, including systems that use data and automated decision-making to place excessive pressure on workers or consumers.

What Does This Mean for the Global AI Industry?

The agreement reflects a growing industry focus on AI agent governance.

As different companies develop their own agents, the ability to connect these systems safely will become increasingly important. Shared standards could make it easier for agents from different platforms to exchange information and complete tasks together.

The agreement also highlights a broader challenge for the AI industry: innovation needs to develop alongside security, privacy and accountability.

However, the convention is a voluntary industry self-regulation initiative. Its real impact will depend on how consistently the participating organisations implement its commitments and how effectively progress is monitored.

The Future of AI Agents Will Depend on Trust

AI agents could become an important part of how people and businesses interact with technology. They may help automate complex workflows, improve productivity and connect services across industries.

China’s new agreement brings major technology organisations together around a common set of principles for developing these systems.

The next challenge will be turning those principles into practical safeguards, technical standards and measurable results.

As AI agents become more independent and widely used, building trust will be just as important as improving their capabilities.