In a meeting with journalists yesterday, Facebook detailed changes it is making to its family of products aimed at dealing with what it calls ‘problematic content.’ For Instagram, this change means a demotion of content Facebook refers to as ‘inappropriate,’ though the posts don’t violate the platform’s Community Guidelines.
Demoted content will not appear in Instagram’s Explore or hashtag pages, according to a new page on the platform’s Help Center. The demotion applies to posts that ‘might not be appropriate for our global community,’ even if those posts aren’t in violation of the Community Guidelines, the company explains.
Instagram’s Help Center page, seen in the screenshot below, doesn’t offer any visual examples of ‘inappropriate’ content, only providing ‘sexually suggestive’ as one category that will be demoted. This change doesn’t apply to the user Feed at this time, but there’s no word on whether Facebook will lower the Feed ranking of these posts in the future.
TechCrunch has published multiple images from Facebook’s press event that include visual examples of ‘non-recommendable’ content set for demotion on Instagram. Though dealing with certain posts, such as ‘likes’ spam and fake news, would obviously be a good thing for users, other categories encompass large, vague content segments with no clear definition of what is and isn’t ‘appropriate.’
Based on the images from Facebook’s press event, Instagram will demote posts that feature sexually suggestive, ‘graphic/shocking,’ and violent content. Examples include, among other things, images of someone being sprayed with pepper spray, a woman in a bikini and a skull.
The vague nature of Facebook’s sweeping ‘non-recommendable’ categories leaves many users in a state of uncertainty and may reduce the platform’s overall usefulness for certain creatives and brands. Though a user’s existing followers will still see the ‘inappropriate’ posts in their respective Feeds, being filtered from Explore and hashtag pages greatly limits the user’s ability to gain new followers.
Though Facebook didn’t go into details in its press release about the changes, TechCrunch reports that the company will use AI algorithms to determine which posts are demoted. Human content moderators have been tasked with labelling ‘non-recommendable’ content; Instagram will use those labels to train machine learning algorithms on identifying borderline content.
Instagram’s content demotion effort is now underway.