Discord says an AI moderation bug wrongly banned thousands of users

TechCrunch reports that Discord admitted an AI moderation bug wrongfully banned more than 8,000 users over a two-month period after harmless images were mistaken for violations. The company said human review was meant to be part of the process, but the bug immediately banned affected accounts. Discord also said roughly 200 more users were wrongly banned over the weekend and that accounts are being restored. This is a factual trust-and-safety failure, not just generic AI backlash.

Key details: Discord said more than 8,000 users were wrongly banned over two months, The bug involved harmless images flagged by AI moderation, The company said accounts are being restored, About 200 additional wrongful bans occurred over the weekend.

Why it matters: At consumer-platform scale, AI moderation failures can become mass account-action events before users get meaningful recourse.