Got my Twitter account suspended for quoting a politician

I was sitting in my truck waiting for a pickup last Tuesday and shared a direct quote from a state senator about voting laws. Within 2 hours my account was locked for 'targeted harassment' even though I never added my own opinion. It took me filing 3 appeals with screenshots of the original source before they reinstated it 4 days later. Has anyone else run into this where platform bots can't tell a quote from an attack?

3 comments

3 Comments

aaronsullivan13d ago

Oh man, that's frustrating. But I gotta gently push back on one thing - you said the bots can't tell a quote from an attack, but actually the bot doesn't care about the difference. It's looking for patterns like flagged keywords or certain user reports, not the context of what you wrote. I've seen people get hit for posting dictionary definitions of words that got flagged by the system. The bot just sees "this person used these words" and then it checks if enough people reported it or if the word combo trips some alarm. It's not smart enough to read your actual intent or see the source link. So yeah, your quote looked exactly like an attack to the algorithm because it matched the same language pattern that real harassers use.

charles_young9212d ago

yeah thats a good point actually, i read somewhere that these moderation bots are basically just using keyword matching and report thresholds. there was this whole thread on a tech forum where a guy tested it by posting a line from a famous book that had a slur in it and he got hit with a warning in like 30 seconds. the bot didnt care that it was a quote, it just saw the word and triggered. kinda scary how dumb the system is but also makes you realize why some people get caught up in false flags so often. i mean if the algorithm cant tell the difference between a dictionary entry and someone actually attacking someone else, thats a pretty big flaw. but i guess the companies figure a few wrong bans are better than letting real harassment slip through.

the_susan12d agoTop Commenter

Is it really that big of a deal though, @aaronsullivan? I mean, yeah, the bot is dumb but getting a warning isn't the end of the world and you can just appeal it half the time. Plus if the alternative is letting real trolls run wild I can see why they lean the way they do.