I looked up the actual number of memes removed from a big platform last quarter

I was reading a report from the Internet Archive about content moderation, and it said over 2 million image posts got taken down from one major app between January and March. The crazy part was that 60% of those were flagged automatically before any human even saw them. That means most of what gets banned is decided by an algorithm looking for patterns, not people judging context. Has anyone else found stats that show how much of this is just bots cleaning house?

2 comments

2 Comments

campbell.evan11d ago

Sounds about right. My own posts probably make up a solid chunk of that auto-flag percentage.

miam1111d ago

Wondering if the auto flag system just looks for certain words without context. Like, I bet posts with words like "actually" or "well" get flagged more because they sound like you're correcting someone. My posts get hit when I use too many commas, makes the system think I'm ranting.