On 12 December 2023 the Central Cyberspace Affairs Commission issued a notice launching a one-month campaign against harmful content orientation on short-video platforms. The notice identifies three categories: false information, inappropriate behaviour and misleading values.
Compared with simple deletion, the campaign focuses on “orientation”. Platforms must judge not only whether content is unlawful but whether it spreads wrong values, behaviour models or misinformation.
What the three categories cover
First, false information: fabricated social events, staged scenes, splicing, old videos presented as new events. Such content often relies on emotional headlines and algorithmic recommendation, so platforms must address both content and distribution.
Second, inappropriate behaviour: dangerous stunts, vulgar performance, flaunting wealth, inducing consumption and risks of imitation by minors. Platforms must assess whether the behaviour itself sets a risky example, not only whether the image breaches rules.
Third, misleading values: spreading anxiety, manufacturing conflict and promoting distorted values. This is the hardest to detect and requires semantic understanding plus human review.
Three lines of remediation
First, content. Build review rules and sample libraries for the three categories, specifying when recommendation is restricted, when content is removed and when accounts are banned.
Second, algorithms. The problem is often not a single video but a recommendation mechanism that amplifies harmful content. Platforms should review objective functions, traffic allocation and intervention in trending topics, avoiding “the more controversial, the more recommended”.
Third, accounts. Accounts repeatedly posting harmful content need account-level action: throttling, suspension, removal of monetisation rights and bans. MCN agencies should be brought into scope, with responsibility for signed accounts.
Minors
The impact on minors is a priority. Platforms should check whether minors’ mode genuinely reduces harmful content, whether dangerous content invites imitation, and whether tipping, top-ups and live-stream connections are restricted for minors.
After the campaign
The campaign has a deadline, but the rules do not disappear with it. Platforms should embed the review standards, sample libraries and handling procedures developed during the campaign into daily governance, avoiding a return to laxity once the campaign ends.
For platform content governance and algorithm compliance, see our algorithm filing service; for user information and minors protection, see data compliance; for platform rules and licensing, see business licensing.
This article is based on the notice on the Qinglang campaign against harmful short-video content orientation issued by the Central Cyberspace Affairs Commission on 12 December 2023. It is for reference only and does not constitute legal advice.