WDTA Releases First International Standard for Large‑Model Supply‑Chain Security

During the 2024 Inclusion·Bund Conference, the World Digital Technology Academy (WDTA) officially released the international standard "Large Model Supply‑Chain Security Requirements," developed jointly by experts from the Cloud Security Alliance (CSA) Greater China, Ant Group, Microsoft, Google, Meta, PrivateAI and dozens of other domestic and international organizations.

This is the industry’s first international standard dedicated to the security of large‑model supply chains, representing a solid advance in global AI governance cooperation.

Peter Major, Executive Chair of the UN Committee on Science and Technology for Development and Founder of WDTA, emphasized that as AI continues to impact all sectors, international cooperation on AI standards becomes increasingly vital for building trust, fostering innovation, and ensuring reliable operation of AI systems worldwide.

Li Yuhang, Chair of CSA Greater China, noted that large‑model systems rely on a globally interconnected supply‑chain ecosystem, presenting unprecedented opportunities and significant security challenges; the new standard provides a systematic management framework for the entire lifecycle of large models, addressing supply‑chain risks and supporting safe, sustainable AI development.

The standard, part of WDTA’s AI STR (AI Safety, Trustworthiness, Responsibility) series, outlines a security protection framework covering data preparation, model development, deployment, and operation, identifying risks such as data leakage, model tampering, and non‑compliant suppliers, and offering guidance for risk assessment and mitigation.

It also serves as a reliable basis for third‑party audits and regulatory compliance, enhancing overall trustworthiness of large‑model systems.

WDTA, founded in April 2023 as an international non‑governmental organization registered in Geneva, promotes digital technology worldwide and leads the AI STR initiative, with members including Microsoft, Google, Anthropic, Ant Group, Baidu and others.

At the recent UN Tech Conference, WDTA released two additional large‑model security standards co‑authored by OpenAI, Ant Group, iFlytek, Google and others, further expanding the AI STR portfolio.

In the context of heightened global focus on AI development and governance, China is also advancing its AI governance framework, issuing policies such as the "Global AI Governance Initiative" and the "National New‑Generation AI Standard System Construction Guide".

Forum speakers shared perspectives on fostering global consensus for AI safety, including Min Jinghua, Deputy Leader of WG7 of the National Cybersecurity Standardization Technical Committee, who highlighted the essential role of technical standards in balancing AI value and risk, and Wang Weiqiang, General Manager of Ant Group’s Machine Intelligence Department, who described Ant Group’s early adoption of trustworthy AI technologies and its integrated large‑model security solution “Ant‑Sky‑Jian”.

The event also launched the "AI Trustworthiness Analysis" report, compiled by CSA Greater China members such as China Telecom, Xidian University, Guangzhou University, Ant Group, Baidu and Tianyi Cloud, which evaluates AI reliability, safety, transparency and fairness, and proposes strategies, technical measures, policy recommendations and industry best practices.