Instagram management informed about new measures to reduce visibility “potentially harmful” content in the app. The algorithm that ranks posts in the main feed and stories will now deprioritize that content “May contain bullying, hate or incitement to violence”.
The company noted that the rules for using the service already contain a ban on posting such materials, however, new measures would be justified in controversial or borderline cases, as well as when posting content that has not yet been deleted by moderators. The system works fully automatically and numerous criteria, including captions, are taken into account for the analysis of the content and its assignment to a specific category. For example, if an image was previously deleted and the associated caption appeared in a new publication, such a publication is immediately pessimized.
Previously, the platform only ranked content in the public Explore section, but now the service has decided to take care of the content that users see right in their feeds when the authors they follow post something inappropriate. Instagram started doing something similar in 2020 — then pessimized over accounts condemned by moderators for disinformation. The platform’s new feature does not involve downgrading accounts as a whole, only individual records.
The administration of the service also said that the new algorithm takes into account the history of complaints submitted by users. If for some reason the system decides that the user can also complain about the next release, their position in the feed will be downgraded.