Breaking

Discord Says AI Moderation Bug Wrongly Banned Over 8,000 Users

July 7, 2026 at 15:32 EDT

Security
Sales & Business

Discord has acknowledged that a bug in its AI-assisted moderation system misflagged harmless images as harmful content, wrongfully banning more than 8,000 users over the past two months. All affected accounts are now in the process of being restored.

July 7, 2026 · Discord

A Moderation Bug Wrongfully Banned Users Over Harmless Images

Discord admits its child-safety detection system instantly banned thousands of accounts for uploading benign "grid-like" images — chess boards, spreadsheets, game inventories — and says all affected users are being restored.

~8,200

accounts wrongfully banned in total

~200

hit specifically by grid-like images

100%

of affected accounts being unbanned

Who was banned since May

Most wrongful bans came from other benign images — grid-like images were a smaller slice.

~8,000

other benign images

~200

grid-like images

How one image became a ban

Upload

Chess board, spreadsheet, game texture, plain background

→

Hash match

False positive against known harmful content

→

Instant ban

Bug banned instead of pausing for manual review

→

No reversal

Second bug failed to auto-restore after review

The cause was not "AI" itself but false-positive hash matches, later fixed by overriding the problematic hashes.

Credit given

Some users praised the fast fix and public apology. Discord says it is developing better safeguards.

Strong backlash

Developers and community operators called losing an account devastating for work, and criticized slow support. The core need: more accurate hash matching.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.