Discord’s help crew took to X yesterday to explain how a bug in its moderation instruments brought on accounts to be incorrectly banned. Whereas about 200 accounts had been caught by the problem over this weekend, the issue could have impacted round 8,200 accounts since Might 2026.
Posts of sq. grid photographs appeared to be the problem and had been mistakenly flagged by Discord as little one sexual abuse materials. That sadly meant something from a spreadsheet to a chessboard to the Minecraft stock menu might need resulted in an misguided ban.
We all know that is not a satisfying rationalization if this was your account, and we should always have caught this sooner. We’re engaged on higher safeguards so this could’t quietly occur once more, and extra broadly, on ensuring our security techniques do not penalize individuals who did nothing fallacious….
— Discord Help (@discord_support) July 7, 2026
The corporate makes use of AI-powered content material moderation instruments to match recognized dangerous materials towards what its members submit. The help crew acknowledged that the system is understood to create false positives, so often flagged accounts are reviewed by an individual from the corporate’s Belief & Security crew earlier than any motion is taken towards a consumer. “The meant conduct is to briefly pause uploads throughout that overview, not ban the account,” the posts state. “We had a bug that brought on the latter. When our workers reviewed and cleared these accounts, the identical bug prevented the ban from being lifted mechanically, so it simply stayed in place.”
As Discord says in its apology, it is an embarrassing mistake to have energetic for roughly two months. The corporate mentioned any accounts that had been banned as false positives ought to now be reinstated.
