Discord: AI moderation bug wrongfully banned 8,000+ users

Original: Discord admits AI moderation bug wrongfully banned users over harmless images

Why This Matters

The incident spotlights growing reliability concerns as platforms scale AI-driven content moderation.

Discord confirmed a bug in its AI moderation system wrongfully banned over 8,000 users since May 2026. Harmless images — including spreadsheets, chessboards, and game textures — were incorrectly flagged as harmful content. All affected accounts are being restored.

Discord has acknowledged that a bug in its automated safety system mistakenly banned more than 8,000 users over a two-month period beginning in May 2026. The system, which works by matching uploaded content against databases of known harmful material, incorrectly flagged benign images — including spreadsheets, chessboards, game textures, and white or gray transparent backgrounds — as harmful content. An additional 200 users were banned over the weekend before Discord's team identified and fixed the issue. All affected accounts are currently being restored.

According to Discord, a human Trust & Safety team member is intended to review flagged content before any action is taken, but a bug caused accounts to be immediately banned without that review. 'We're working on better safeguards so this can't happen again,' the company wrote on X.

Users on X and Reddit reported being permanently suspended after uploading images with square grid patterns. Some speculated the AI had become overly sensitive to grid-like visuals due to their historical use in disguising prohibited content. Affected users, including a game developer who relied on Discord for professional communication, expressed significant frustration. Discord is not alone in this challenge — Meta platforms Instagram and Facebook faced similar widespread unexplained suspensions attributed to AI moderation last year.

Source

techcrunch.com — Read original →