11-017 · EVIDENCE B · SOURCE-SENSITIVE
Before offering generative AI services to the public, conduct required safety assessments and algorithm registration; mark all generated content
对公众提供生成式 AI 服务前先按规定做安全评估和算法备案，生成内容加标识
There is no monetary cost, but you must follow procedures set by internet regulators — completing them takes some time. Small personal tools that are not made available to the public do not fall under these rules.
If you offer generative AI services with public influence or social mobilization potential to the general public, you must perform a safety assessment and register the algorithm per national regulations before launch. Any changes or shutdowns also require filing updates or cancellation paperwork. All generated images and videos must be clearly marked. Tools meant solely for personal use and not open to the public are exempt from these requirements. The regulation itself does not specify penalties; violations are handled under the Cybersecurity Law, Data Security Law, and Personal Information Protection Law.
B
Move “Before offering generative AI services to the public, conduct required safety assessments and algorithm registration; mark all generated content” to Future Me's calendar
The excuse
Future Me has a clearer calendar, more patience, and apparently the same responsibilities.
What feels good now
Today I keep a few minutes and avoid interrupting my default routine.
What the source says
The source summarizes the return this way: “If you offer generative AI services with public influence or social mobilization potential to the general public, you must perform a safety assessment and register the algorithm per national regulations before launch. Any changes or shutdowns also require filing updates or cancellation paperwork. All generated images and videos must be clearly marked. Tools meant solely for personal use and not open to the public are exempt from these requirements. The regulation itself does not specify penalties; violations are handled under the Cybersecurity Law, Data Security Law, and Personal Information Protection Law.” Evidence grade B. The original states the population, caveats, and sources; this is not a personal forecast.
Use the original facts and limits. For an urgent medical, safety, financial, or legal issue, contact a qualified local professional or service.
THE ENGLISH EDITION
Original recommendation and evidence
17. Before offering generative AI services to the public, conduct required safety assessments and algorithm registration; mark all generated content
Cost: There is no monetary cost, but you must follow procedures set by internet regulators — completing them takes some time. Small personal tools that are not made available to the public do not fall under these rules.
In plain terms: If you offer generative AI services with public influence or social mobilization potential to the general public, you must perform a safety assessment and register the algorithm per national regulations before launch. Any changes or shutdowns also require filing updates or cancellation paperwork. All generated images and videos must be clearly marked. Tools meant solely for personal use and not open to the public are exempt from these requirements. The regulation itself does not specify penalties; violations are handled under the Cybersecurity Law, Data Security Law, and Personal Information Protection Law.
Benefit: Providers of generative AI services that can shape public opinion or mobilize people must conduct safety assessments, register their algorithms, and handle subsequent updates or cancellations as required by law. “Public influence” or “social mobilization potential” means the ability to affect how people think or to organize them. Providers must also add identifying marks to any generated images or videos. This regulation took effect nationwide on August 15, 2023.
Evidence grade: B
Sources: Cyberspace Administration of China and seven other departments (2023). Interim Measures for the Management of Generative Artificial Intelligence Services (Articles 12, 17). <https://www.gov.cn/zhengce/zhengceku/202307/content_6891752.htm>
Notes: The regulation itself does not state exact penalty amounts. Violations are penalized under the Cybersecurity Law, Data Security Law, and Personal Information Protection Law, which is why this item is rated B. No civil cases involving copyright or portrait rights violations caused by AI‑generated content are cited here, as the original texts were not verified on the Supreme People’s Court website.
Read the complete Chinese source and notes
17. 对公众提供生成式 AI 服务前先按规定做安全评估和算法备案，生成内容加标识
成本: 不花钱，但要走网信部门的流程，办下来要花些时间。个人做的小工具只要不对公众开放，就不触发这一套
说人话: 面向公众、能影响舆论或动员人的生成式 AI 服务，上线前要做安全评估、办算法备案，改了或停了还要办变更、注销。生成的图片和视频要加标识。自己用、不对公众开放的小工具不用管这些。违反了按网络安全法等法律处罚。
收益: 提供具有舆论属性或者社会动员能力的生成式人工智能服务，应当按国家规定开展安全评估，并按算法推荐管理规定办算法备案，以及后续的变更、注销手续。舆论属性或者社会动员能力，指的是能影响公众看法，或者能把人组织动员起来。提供者应当对图片、视频等生成内容进行标识。办法自 2023 年 8 月 15 日施行（全国）
证据等级: B
来源: 国家互联网信息办公室等七部门 (2023). 生成式人工智能服务管理暂行办法（第十二、十七条）. <https://www.gov.cn/zhengce/zhengceku/202307/content_6891752.htm>
备注: 这个办法自己没写罚多少钱。违反了要回到网络安全法、数据安全法、个人信息保护法里找处罚，所以定 B 级。用 AI 生成的内容侵犯别人著作权、肖像的民事案例，本节未在最高法官网核实原文，不引。低价转卖境外大模型接口的中转站，见本节第 19 条（AI 中转站）。
YOUR PRIVATE LIST