À suivreGuide suivant
China’s Algorithm Registry and Large Model Filing System
Société
GUIDE DE LA SOCIÉTÉ
China’s Provisions on the Administration of Algorithm-generated Recommendations for Internet Information Services took effect on March 1, 2022.
They cover recommendation services such as ranking, selection, pushing, and retrieval, with duties that vary by provider and service. The rules address user controls, worker protections, content governance, transparency, and algorithmic price discrimination.
The Cyberspace Administration of China and three other agencies issued the Provisions on the Administration of Algorithm-generated Recommendations for Internet Information Services. They took effect on March 1, 2022. The rules reach internet information service providers that use algorithmic recommendation technology to provide services, including synthetic content, personalized push, ranking, search filtering, and work scheduling. Coverage turns on the specified service and conduct, not any use of machine learning. The provisions require providers to inform users in a conspicuous way that recommendation services are being used and disclose basic principles, purposes, and main operating mechanisms appropriately. They require an option that does not target a user’s personal characteristics or a convenient way to turn off recommendation services; after a user opts out, the related service must stop. Users must also be able to select or delete personal tags used for recommendation. The provisions address minors and older people, and impose additional governance on providers with public-opinion or social-mobilization capacity. Work-related rules apply to platforms using recommendation algorithms to schedule work or determine compensation. They call for improving dispatch, compensation, working hours, and rest protections, and for engaging workers’ representatives or unions in relevant arrangements. They prohibit unreasonable differential treatment in transaction conditions, including price, based on consumer preferences, transaction habits, or other characteristics without legitimate justification. This is the context for the colloquial phrase “big data killing the familiar,” but the legal test should be taken from the text rather than the slogan. Providers have further obligations concerning algorithmic security, records, assessment, and cooperation with regulators. Some providers must complete algorithm filing with the CAC and publish filing information. This is a specific administrative filing regime; it is not a general license required of every AI model. The provisions sit within China’s broader internet-content and data-governance system. A company assessing exposure should examine service function, user location, provider role, filing status, and later rules that may also apply.
Les dommages catastrophiques et quotidiens causés par l’IA dépendent tous deux de la personne qui comprend les risques et qui peut agir.
Les connaissances du public et des professionnels déterminent si une politique de sécurité forte est politiquement possible.
Des explications claires réduisent la capture par le battage médiatique, les relations publiques en laboratoire et le théâtre d'éthique vague.
China continues to develop AI and online-service rules that may overlap with the 2022 recommendation provisions. The CAC published a July 2026 draft revision to the Internet Information Service Management Measures for public comment; treat it as a proposal unless and until a final measure is adopted. Monitor official CAC notices, preserve the controlling Chinese text, and obtain jurisdiction-specific advice when operating in China. Machine translations may miss defined terms or filing triggers. Recheck whether the draft was finalized before a material launch or service change, and assess any new measure alongside the recommendation rules.
A covered app offers a clear option to turn off recommendation services or choose a feed not based on personal characteristics.
A platform gives users a way to manage or delete interest labels used to shape personalized content.
A delivery platform assesses dispatch and compensation rules for effects on couriers’ rest, pay, and workload.
A service reviews whether it uses different transaction prices for comparable products based on a user’s personal characteristics.
Traiter le risque existentiel comme de la science-fiction alors que les capacités s’accroissent.
Confondre sécurité des produits de surface et alignement sous haute autonomie.
Laisser le public non anglophone et non expert avec uniquement des sources de mauvaise qualité.
Séparez les dommages causés aux produits, leur mauvaise utilisation et les risques de perte de contrôle/désalignement.
Demandez quelles preuves pourraient changer votre point de vue sur les délais et la gravité.
Préférez les sources primaires et les évaluations concrètes aux allégations marketing.
Identifiez une voie d’action : carrière, politique, financement ou compétences – et pas seulement la sensibilisation.
Free newsletter
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
China’s Provisions on the Administration of Algorithm-generated Recommendations for Internet Information Services took effect on March 1, 2022. They cover recommendation services such as ranking, selection, pushing, and retrieval, with duties that vary by provider and service. The rules address user controls, worker protections, content governance, transparency, and algorithmic price discrimination.
The official CAC announcement states that the provisions took effect on March 1, 2022.
The rules include personalized push, sorting and selection, search filtering, and other listed recommendation functions.
The provisions require one of these user controls and require the related service to stop after opt-out.
The provision says the provider must immediately stop providing the related service after the user chooses to turn it off.
The provisions address algorithmic scheduling and compensation arrangements affecting workers’ rights and interests.
Continuez à apprendre
Plus de guides sélectionnés pour ce sujet
À suivreGuide suivant
China’s Algorithm Registry and Large Model Filing System
Société