AI Answer Library

What Chinese compliance requirements apply when a company uses large language models?

Short answer

First determine whether you are using AI internally or offering a generative AI service to the public in mainland China — that single distinction sets the scale of your compliance burden. China's Interim Measures for the Administration of Generative AI Services primarily bind parties that provide such services to the public; enterprises that develop and apply the technology without offering a public-facing service fall outside their scope. The Data Security Law, the Personal Information Protection Law and the cybersecurity classified protection scheme, by contrast, apply to everyone. In practice, classify your data first, decide whether it may leave your perimeter second, and only then discuss model selection.

Key points

  • 01Internal use and providing a public-facing service are two entirely different compliance tracks. The Interim Measures state explicitly that entities which develop and apply generative AI without offering a service to the public in mainland China fall outside their scope.
  • 02When you do serve the public, obligations extend beyond content safety and authenticity to algorithm filing, with an additional security assessment for services carrying public-opinion or social-mobilisation attributes. Bring legal counsel in at the product design stage for this track.
  • 03Personal information processing is governed by the Personal Information Protection Law regardless of track: notice and consent, data minimisation, and separate consent for sensitive personal information. Clear that gate before bulk-loading customer records into a knowledge base.
  • 04The Data Security Law requires classified and tiered data protection. Its practical effect on an AI project is direct: it determines which corpora may enter a knowledge base, which knowledge base, and whether an external API may be called at all.
  • 05Cross-border data transfer is a separate topic, not merely "calling an API". Outbound transfers run through distinct routes — security assessment, standard contract, or certification — and which applies depends on the type and volume of data. Assess it on its own before signing.
  • 06Publicly presented AI-generated content carries labelling obligations. If your product shows model-generated text, images or audio-visual material to external users, build labelling into the design rather than retrofitting it after launch.

The main instruments and what each one governs

The table maps the instruments a mainland-China enterprise AI project most often touches, split into what each governs, who it binds, and the practical implication — enough to judge at kickoff which ones concern you. Two caveats: this is directional only, cites no article numbers, and does not substitute for legal advice. Sector-specific rules commonly layer on top in healthcare, finance, education and industry, and regulatory interpretation evolves with practice. The binding conclusion must come from your own counsel, applied to your specific business.

Instrument / requirementWhat it governsWho it bindsPractical implication
Interim Measures for the Administration of Generative AI ServicesThe act of providing a generative AI service: content safety, lawful training data, labelling, user managementParties offering generative AI services to the public in mainland China; development and application without a public offering is out of scopeDecide which category you are in first. If you will serve the public, put algorithm filing and security assessment into the project plan
Data Security LawClassified and tiered data protection, and the handling obligations attached to important dataEvery organisation carrying out data processing activities within mainland ChinaGrade the corpus first, then decide what may enter a knowledge base and whether it may leave the network
Personal Information Protection LawNotice and consent, minimisation, separate consent for sensitive data, and responding to individual rights requestsEvery organisation processing personal information, internal systems includedRedact or obtain authorisation before ingesting customer records, and keep deletion and export technically feasible
Cybersecurity classified protection (MLPS 2.0)Security controls and management measures by system grade, plus assessmentNetwork operators in mainland China; the grade follows system criticality and impactAn AI system is usually graded together with the business system it serves — budget the assessment cycle into the schedule
Cross-border data transfer rulesThe routes and preconditions for sending data abroad: security assessment, standard contract, or certificationAny processor that needs to transfer data outside mainland ChinaCalling an overseas model API can itself constitute an outbound transfer — assess it separately rather than assuming it is fine
Labelling of AI-generated contentAI-generated or synthesised content shown to the public must be identifiableProducts that present generated content directly to external usersBuild labelling in at design time; retrofitting after launch is expensive

A workable order for making the compliance call

Question one: will this system offer a generative AI service to the general public in mainland China? If yes, the compliance path becomes materially heavier — algorithm filing, content safety mechanisms and user management all need advance planning, with legal counsel engaged at design time. If no (internal staff only, or contracted customers in a controlled environment), the centre of gravity shifts to data and security. Question two: what data will it touch? Following the Data Security Law's tiering logic, sort the corpus into three bands — publicly available, internal-general, and sensitive or personal — and state which systems each band may enter. Question three: will data leave the internal network, or the country? Calling an overseas model API may constitute an outbound transfer, and the answer here often decides self-hosting versus API on its own. Question four: what protection grade applies and when is the assessment? Put the assessment cycle in the schedule rather than the week before launch. Walking these four in order resolves most disputes while the project is still on paper.

Technical capabilities the compliance story depends on

Compliance is not only paperwork; it imposes hard requirements on the system, and without them you cannot demonstrate compliance at all. First, complete audit logs: who asked what and when, which documents were retrieved, what the model returned — all traceable and retained for the required period. Second, permission isolation: retrieval must filter by user and department, or anyone can extract material they are not cleared to see simply by asking. Third, deletable and exportable data: individuals may demand erasure of their personal information, and if the knowledge base cannot locate and remove one person's data, that obligation cannot be met. Fourth, filtering and refusal: requests outside the authorised scope must be declined rather than answered anyway. Fifth, the ability to switch things off: take a capability or a corpus offline quickly when something goes wrong. All five are straightforward under self-hosting; on an architecture built directly on external APIs they must be added at the gateway layer.

Where this applies

When this answer does not hold

  • This is not legal advice. It offers directional orientation only, cites no article numbers, and the binding interpretation rests with the competent authorities and with your own in-house or external counsel.
  • Rules and enforcement practice keep evolving; this reflects publicly available information as of August 2026. On a project spanning several months, re-check the current requirements at kickoff and again before launch.
  • Sector-specific rules apply on top. Regulators in healthcare, finance, education, automotive and industrial control typically impose their own data and algorithm requirements, and this general overview does not replace them.
  • This covers the mainland-China context only. Where Hong Kong, Macau, Taiwan or overseas users and deployments are involved, the local regimes must be assessed in parallel and may lead to different conclusions.

People also ask

  • Does a company need to file a registration to use AI in China?
  • What does the Interim Measures for Generative AI Services require of companies?
  • Does internal LLM use count as providing a service to the public?
  • Does an AI system need to go through China's cybersecurity classified protection scheme?
  • What should we watch for when an LLM processes customer personal information?
Written by: YGG Technology Solutions TeamPublished: 2026-08-01Last reviewed: 2026-08-01