Discord Fixes AI Moderation Bug After Harmless Images Trigger Wrongful User Bans

Discord has confirmed that a flaw in its AI-powered moderation system mistakenly banned more than 8,000 users during the past two months after harmless images—including spreadsheets, chessboards, game textures, and plain white or grey transparent backgrounds—were incorrectly identified as harmful content.

The company said the issue had been affecting user accounts since May. Before engineers discovered and resolved the bug over the weekend, roughly 200 additional users were also suspended. Discord stated that it is now restoring all impacted accounts as the recovery process continues.

The incident underscores a growing concern surrounding AI-assisted moderation, as more online platforms rely on automated tools to detect illegal, abusive, or harmful material across large volumes of content.

In a detailed post on X, Discord explained that its automated safety technology compares uploaded files against databases containing known harmful material. Although the system is intended to identify illegal content, the company acknowledged that it can occasionally produce false positives. While flagged content is normally reviewed by a human moderator, a software bug caused affected accounts to be permanently banned immediately instead.

“We’re working on better safeguards so this can’t happen again,” Discord said in its statement.

On both X and Reddit, numerous users reported being permanently suspended after uploading images containing simple square grid patterns. Some users suggested that Discord’s AI moderation system may have become increasingly sensitive to grid-like designs because similar patterns have previously been used to conceal NSFW material and child exploitation content from automated detection tools.

Many affected users have voiced their frustration on social media, arguing that permanent account suspensions triggered solely by automated systems can have serious consequences, especially for people who rely on Discord for work, online gaming communities, or staying connected with friends and family over long distances.

“Losing a Discord account to something as unfair as this can be extremely devastating and affect users severely, and every day, millions of users are affected by false AI bans. This needs to be stopped,” one user wrote on X.

Discord is not the only technology platform to encounter problems related to automated moderation. Last year, users of Instagram and Facebook Groups reported a wave of unexplained account suspensions that many believed were caused by AI moderation systems. Although many users blamed automation, Meta never publicly confirmed whether AI-related errors were responsible. Since then, Meta’s Oversight Board has called for greater transparency around the company’s moderation practices.

Tumblr also faced similar criticism last year after users reported that large numbers of accounts had been suspended without receiving clear explanations for the actions taken against them.