If you rank or recommend to the public in China, filing is a CAC process — not a privacy-policy paragraph.
CAC rules on recommendation algorithms require many public-facing ranking, recommendation and similar engines to file, publish user-control mechanisms, and meet content and labour-style fairness duties. Filing is not PIPL consent and not ISO 27001. GAI services have a related but distinct security-assessment/filing track (sibling). The live consumer-digital compliance article remains the long guide — this wiki page is the filing concept. Portal fields and catalogue codes change; they are not reprinted here. Internal HR ranking tools are primarily PIPL ADM, not this public-filing regime.
4 questions before you choose the route.
This page identifies the right question and evidence. It does not determine the legal outcome on a reader’s facts.
Is the algorithm public-facing in China?
Feed, search, recommend, price display.
PublicWhat type is it under current CAC lists?
Recommendation vs GAI vs other.
TypeHave user controls been built?
Opt-out of targeting, explanation.
ControlsIs PIPL ADM also in play?
Scoring people.
PIPLWorking rule: Map the regulated role before marketing or launch in China.
The signal ledger.
These facts move the question beyond a label and into a product, money-flow and control analysis.
Bring a compact evidence docket—not a pitch deck.
Give a compliance team or counsel the operating facts that reveal the perimeter.
Questions people ask before they build.
Short answers for orientation. The right result can change with the service model and current rules.
Does every ML model need filing?
No. The regime targets specified public information-service algorithms. An internal demand-forecast model is a different story.
Where is the long guide?
Keep using /china-consumer-digital-compliance-algorithm-pipl for the national narrative. This page is the atomic filing definition.
Primary authorities
Reviewed sources support orientation, not a fact-specific assessment.