Every AI Policy Ever · open in the interactive register
Published (non-binding, current)Non-binding · Guidelines / framework
网络安全标准实践指南——个人用户使用人工智能服务安全指南
Record CN-27 · TC260-PG-20267A
TC260 practice guide on how individual users can use AI services over the internet securely. Users should recognise AI's limits such as errors and hallucinations, avoid over-reliance and replacing real social contact, and watch for misuse such as cloned voices of relatives or forged official notices. They should use services from regular channels and generative AI built on filed large models, check agreements and permissions, protect sensitive information, cross-check outputs, treat legal, medical or financial outputs as reference only, and not induce harmful outputs, impersonate others or remove AI content labels.
No English version of this instrument was found. The English title and the summary are the register's own, written from the sources below; for the exact wording use the Chinese text.
Version v1.0-202608, released with notice 网安秘字〔2026〕105号 of 24 Aug 2026. It also serves as a reference for technology providers, developers and operators of AI services. Guardians of minors and elderly users are asked to help them use AI services safely. For deploying agents it refers users to TC260-PG-20266A, and for the style of content labels to chapter 5 of GB 45438-2025 (CN-08).
Open the official source https://www.tc260.org.cn/tc260/sjzn/202608/6bd7f30173444f6cae692e0e7c3045b8/files/23e68462a22b465c9aef09e584659924.pdf
Secondary source: https://www.tc260.org.cn/tc260/tzgg/202608/1586c71e2ebf4db28ed4a7651788ca87.shtml
Archived. Last attempt 2026-10-18 00:45:34 UTC
GB 45438-2025 (AI content labeling standard) China · In force
TC260-PG-20266A (agent deployment and use security) China · Published (non-binding, current)
National Cybersecurity Standardization Technical Committee (TC260) Secretariat. Cybersecurity Standard Practice Guide — Security Guide for Individual Users Using Artificial Intelligence Services [TC260-PG-20267A]. 2026-08-24. https://www.tc260.org.cn/tc260/sjzn/202608/6bd7f30173444f6cae692e0e7c3045b8/files/23e68462a22b465c9aef09e584659924.pdf. Status: Published (non-binding, current); verified 2026-10-18 (level B). Local copy archive/_runs/2026-10-14-discover-leads-2026-10-14/C011/2026-10-14/official.pdf retrieved 2026-10-14 13:11:03 UTC, SHA-256 2b8b471b5a5aa188ab08201913a14d3ddce70a6dccf222e2d34ddb3fff39f014. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record CN-27.
Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.