Baidu’s quality framework treats pirated and copyright-infringing content explicitly, and it appears in the grading examples rather than as a separate policy — which tells you how routine the problem is.
Where it appears in the grading
- Medium quality: a film playback page with no copyright information; a page of a pirated novel scraped from a known fiction site. Baidu grades these as merely medium — they meet a need, they are just not the publisher’s own work.
- Good quality: a licensed film on a proper video site, playing cleanly. The distinction between the two examples is licensing, not content.
The accessibility dimension
Baidu’s treatment of permission walls is where the copyright judgement becomes concrete. A paywalled academic paper with genuine copyright and high production cost is treated as a normally accessible page, because the restriction is within the user’s expectations. A login or payment wall over cheaply collected, non-original, republished material is judged case by case and may be demoted.
The principle is unusually well stated: restriction is acceptable in proportion to what was invested in the content. Charging for your own work is legitimate; charging for somebody else’s is not.
What this means for a publisher
- State copyright and licensing information on media pages. It is one of the observable differences between the good and medium examples.
- Attribute republished material — Baidu’s scraping rules confirm that news and policy content republished with the source stated is not treated as collection.
- Do not paywall content you did not produce.
The exposure beyond search
Copyright in China is enforced through the courts as well as through search, and platform liability has tightened considerably. A site whose model depends on republishing without permission carries a legal risk that dwarfs the ranking one, and no amount of technical work addresses it.
The practical position
Licence what you publish, say so on the page, and attribute what you republish. All three are cheap, all three are visible to the grading framework, and all three are the same thing a lawyer would ask for.
Comment nous l'appliquons
The paywall distinction is the most useful thing on this page: Baidu accepts a restriction proportional to what was invested, so charging for your own work is fine and gating scraped content is not. We advise clients to state licensing plainly on media pages, because it is one of the few observable differences between Baidu’s own good and medium examples.
Services associés