Apps whose AI features run on another company's model now need written authorization from a CAC-registered model owner, or must geoblock those features, before January 31, 2027.
Apple has added a compliance section to App Store Connect, its developer console, that makes every app sold on the App Store in mainland China declare whether it offers AI-generated or synthetic content or features that need a government security assessment. Apps that answer yes must upload Chinese filing numbers, authorization agreements and assessment reports. According to Apple’s own help page, all existing apps on the mainland storefront must have completed and verified declarations by January 31, 2027, to stay available. IT Home first reported the change on November 17, citing developer feedback, and TechNode reported that new apps must supply the information when they are submitted.
The requirement reaches beyond AI developers. Apple says every app on the China mainland store must complete both declarations, answering no if it offers none of the regulated features or if they are disabled or geoblocked in mainland China. For apps that do offer chatbots, image or video generation, or voice and face alteration, the store now checks paperwork that Chinese rules have required for some time.
The strictest test applies to apps built on someone else’s model. An app whose AI features run on another company’s technology must show a usage agreement from the owner of a model or algorithm listed in the official registry of the Cyberspace Administration of China (CAC). Apple has not published a developer news item on the change, and it has not said how many apps are affected.
Apple now checks filing numbers, model licenses and assessment reports
Apple’s help page splits the requirement in two. The first declaration covers AI-generation or synthetic services. Apps offering them must give the approved filing number and registered filing name of their large language model or algorithm. If the AI comes from a third party, the developer uploads an authorization from the technology owner that names the foundation model, its filing number and the developer as the authorized party, and the authorizing company must match the owner listed in the CAC registry.
The second declaration covers features that require a security assessment. Apple’s examples are distributing AI-generated or synthetic content and offering end-to-end encrypted instant messaging. Developers must submit a screenshot from the National Internet Security Management Service Platform (全国互联网安全管理服务平台) showing the assessment passed, with the app and provider names visible, plus a copy of the assessment report bearing the company seal.
An app that passes App Review but fails the compliance check is released everywhere else and stays off the China mainland storefront until the documents are accepted. Apple lists four statuses: Missing Info, Pending Verification, Verification Failed and Active. A corrected declaration is reviewed only when the developer submits its next app version.
China’s 2025 labeling rule put app stores in the checking role
Apple ties the AI declaration to what its help page calls the Measures for Labeling AI-Generated or Synthetic Content (《人工智能生成合成内容标识办法》). The CAC published the measures on March 14, 2025, issued jointly with the Ministry of Industry and Information Technology, the Ministry of Public Security and the National Radio and Television Administration, and they took effect on September 1, 2025. Article 7 requires app distribution platforms, when reviewing an app for listing, to ask the provider whether it offers AI generation services and, if it does, to verify its materials on labeling generated content.
The security assessment draws on two older texts that Apple also cites. The CAC’s interim measures on generative AI services, issued by seven agencies and in force since August 15, 2023, require services with "public opinion attributes or social mobilization capacity" to undergo a security assessment and complete algorithm filing. The CAC’s 2018 rules, the Provisions on the Security Assessment of Internet Information Services with Public Opinion Attributes or Social Mobilization Capabilities (《具有舆论属性或社会动员能力的互联网信息服务安全评估规定》), in force since November 30, 2018, require a self-assessment when a service launches or adds such functions, or when new technology substantially changes its reach.
IT Home linked the AI declaration to the filing and registration duties in the 2023 interim measures and said Apple’s 2026 App Review Guidelines already carried generative AI clauses. In IT Home’s reading, the update turns those clauses into a hard gate at submission and makes the app store responsible for verification.
The January 31 deadline sets the pace for existing China apps
The cutoff gives developers about ten weeks to gather documents that some may not yet hold. A developer using a model without a Chinese filing has two options in Apple’s framework: secure an authorization from a filed model provider, or disable or geoblock the feature in mainland China and answer no.
Apple tells developers with questions to consult the CAC’s regulatory guidance or the security platform. Whether Apple sends notices to apps still listed as Missing Info as the deadline nears, and how many titles leave the China mainland store after January 31, will be the first measure of how strictly the check is enforced.