Huawei, Alibaba join 22 firms in releasing AI agent convention

Huawei, Alibaba and 20 other organisations have jointly released a self-discipline convention on AI agent development (智慧體發展自律公約), calling for rules expression and embedding, compliance testing, and other technologies to ensure agent behaviour complies with laws and public order.

Meanwhile, the move also aims at resisting algorithmic exploitation and illicit activities conducted through AI agents.

The convention aims to curb platforms or companies using AI and big data to monitor, analyse, and constrain workers or users for maximum profit or labour value, such as AI monitoring of delivery workers or ride-hailing drivers, along with AI-driven crimes such as online fraud and grey-market activities such as ticket scalping or traffic fraud.

Unveiled on 4 November at the World Internet Conference Wuzhen Summit, the convention was jointly drafted and released by AI agent ecosystem players including the China Academy of Information and Communications Technology, Tencent, Ant Group, Alibaba, and Douyin, guided by organisations such as the China Federation of Internet Societies, according to Chinese state media People's Daily Online.

The convention proposes eight self-discipline measures: strengthening the AI agent trust system, building AI agent interconnection infrastructure, raising the level of AI agent interoperability, enhancing AI agent security protection capabilities, improving AI agent governance mechanisms, protecting the rights and interests of AI agent users, accelerating AI agent empowerment across industries, and promoting inclusive sharing of AI agent technologies.

The formulation and release of the convention marks an important step toward standardised development in the AI agent industry, helping build consensus, strengthening industry self-discipline, advancing AI agent mutual trust, interconnection, and interoperability, and guiding standardised application and innovative development.

Next, the initiating organisations will take the lead in implementing the commitments of the self-discipline convention and will actively accept oversight from industry organisations and the public.

At the same time, all types of participants in the AI agent ecosystem are welcome to actively contribute to the convention's development, jointly fostering an open, trustworthy, secure, and inclusive AI agent ecosystem.

Related articles:

STB, Alibaba map out Singapore’s hotspots as Chinese travel searches surge
Alibaba mulls HK$80bn HK share placement to fuel AI expansion
Huawei brings holiday vibes to HK streets with Pura 90s series