Source: The Verge
Summary
Discord acknowledged a bug in its AI moderation system that mistakenly banned over 8,000 users in two months. The system incorrectly flagged harmless images, such as spreadsheets, chessboards, and transparent backgrounds, as harmful content. The issue started in May, according to the company.
Our Reading
The launch follows a familiar script.
Discord’s AI moderation system went rogue, banning users for innocuous images. The company says it’s fixing the issue, but not before 8,000+ users got the boot. Because what’s a few thousand false positives when you’re trying to keep the internet safe? This is what happens when “AI-powered” just means “new”.
Author: Evan Null
AI Moderation: The Never-Ending Story
When will we learn that AI isn’t the silver bullet for content moderation? Discord’s latest debacle is just another example of how these systems can go haywire.
The Buggy AI Blues
Who needs actual harm when you can flag a chessboard as harmful content? The absurdity of this situation is only matched by the number of users affected.
May the Odds Be Ever in Your Favor
May is the month when Discord’s AI moderation system started playing a cruel joke on its users. Who knew a simple spreadsheet could be so threatening?
The Usual Suspects
Transparency is key, but not when it comes to transparent backgrounds, apparently. The list of innocuous images flagged by Discord’s AI is a laundry list of “what could possibly go wrong?”
Fixing the Fix
Discord says it’s working on a fix, but the damage is done. Can we trust these AI-powered systems to get it right next time? Ask the 8,000+ users who got banned for having a chessboard on their profile.
