What does CAC AI Content Labelling Measures (国信办通字〔2025〕2号) say about AI in banking?

Published Mar 7, 2025 · Last reviewed Nov 20, 2026

The CAC, MIIT, the Ministry of Public Security and NRTA issued the Measures for Labelling AI-Generated and Synthetic Content under notice 国信办通字〔2025〕2号, dated 7 March 2025 and published on 14 March 2025. The measures took effect on 1 September 2025. They require explicit labels that users can perceive and implicit labels embedded in file data. The official text is in Chinese.

OFFICIAL TEXT: cac.gov.cn ↗ · IN FORCE · CHINA (NFRA / PBOC / CAC)

Checked 2026-11-20 · China (NFRA / PBoC / CAC), cac.gov.cn
DocumentCAC AI Content Labelling Measures (国信办通字〔2025〕2号) — Measures for Labelling Artificial Intelligence Generated and Synthetic Content (人工智能生成合成内容标识办法)
Issued byNational Financial Regulatory Administration (NFRA), People's Bank of China (PBoC) and Cyberspace Administration of China (CAC)
TypeRegulation
StatusIn force
PublishedMar 7, 2025
EffectiveSep 1, 2025
Applies toNetwork information service providers that fall within the algorithmic recommendation, deep synthesis or generative AI rules (Article 2), including financial firms whose public services generate AI text, audio, images or video. Binding; the official text is in Chinese
Official sourcecac.gov.cn ↗
Use casesGenerative & agentic AI · Customer-facing chatbots

What are the key points of CAC AI Content Labelling Measures (国信办通字〔2025〕2号)?

  • Explicit labels are text, sound or graphics added to generated content or interaction interfaces that users can clearly perceive.
  • Implicit labels are technical markers added to the file data of generated content that users do not readily perceive.
  • The notice was issued jointly by the CAC, MIIT, the Ministry of Public Security and the National Radio and Television Administration.
  • The measures took effect on 1 September 2025.
  • For banks, the measures matter where AI-generated content, such as chatbot replies, synthetic voice or marketing images, is provided to the public.

What did CAC AI Content Labelling Measures (国信办通字〔2025〕2号) change for banks?

Labelling of AI-generated content became a binding duty in China from 1 September 2025, adding to the deep synthesis and generative AI rules.

When did China's AI-generated content labelling rules take effect?

The Measures for Labelling AI-Generated and Synthetic Content took effect on 1 September 2025. The notice issuing them, 国信办通字〔2025〕2号, is dated 7 March 2025 and was published by the CAC on 14 March 2025.

What is the difference between explicit and implicit labels under China's rules?

Explicit labels are text, sound or graphics in the content or interface that users can clearly perceive. Implicit labels are technical markers added to the file data of the content that users do not readily notice. The measures define both and require them on AI-generated and synthetic content.

DateDocumentStatus
Dec 27, 2024NFRA Data Security Measures (金规〔2024〕24号) — Measures for Data Security Management of Banking and Insurance Institutions (银行保险机构数据安全管理办法)In force
Jul 10, 2023CAC Generative AI Interim Measures (Order No. 15) — Interim Measures for the Management of Generative Artificial Intelligence Services (生成式人工智能服务管理暂行办法)In force
Jan 10, 2022CBIRC Digital Transformation Opinions (银保监办发〔2022〕2号) — Guiding Opinions on the Digital Transformation of the Banking and Insurance Industries (中国银保监会办公厅关于银行业保险业数字化转型的指导意见)In force
Dec 31, 2021CAC Algorithmic Recommendation Provisions (Order No. 9) — Provisions on the Management of Algorithmic Recommendations in Internet Information Services (互联网信息服务算法推荐管理规定)In force
Dec 29, 2021PBoC FinTech Development Plan 2022-2025 — FinTech Development Plan (2022-2025) (金融科技发展规划(2022-2025年))Final
Nov 19, 2026BNM AI Feedback Statement — Feedback Statement: Artificial Intelligence (AI) in the Malaysian Financial SectorFinal

Follow every document these regulators publish

when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am PT · free

plus every regulation, bank and agent page update, the morning after · leave any morning