User tweeted: "After 3 long months, it’s finally “breakfast before work” again! #danedecor #COVID__19 #backtowork #coronavirus" along with a photo of his breakfast plate with bacon, omelet and English muffin. This image was censored with a sensitive content filter. 

I have been attempting to record a cross section of sensitive content filters that are inappropriate, because something is clearly wrong with their algorithm. I have not yet found a clear pattern, but tweets using certain hashtags seem to be more subject to the inappropriate use of a sensitive content filter. They do not seem to be filtering every image associated with certain hashtags, though, so I'm not sure where the algorithm is failing. The more of these innocuous tweets that we log, the better able we may be to at least make the case that the algorithm is broken, even if we don't see a clear pattern in it. With only 20 followers, this user should not have a picture of his breakfast plate censored, but using the hashtags related to reopening the economy, which has been a controversial topic, seems to have put him on the radar.