Discord admits to an AI management bug that wrongly bans users on harmless images

Discord has admitted that a bug in its AI moderation system mistakenly led to it banning more than 8,000 users over the past two months, after it incorrectly flagged innocuous images — including spreadsheets, chess boards and game textures, as well as transparent white and gray backgrounds — as malicious content.

The company confirmed that the problem had affected accounts since May, with an additional 200 users banned over the weekend before its team identified and fixed the issue. All affected accounts are currently under restoration.

The incident highlights one of the growing challenges surrounding AI-assisted moderation as many platforms increasingly rely on automated systems to identify illegal or offensive material at scale.

In a Detailed topic about XDiscord explained that its automated security system works by matching uploaded content against databases of known malicious material. While this technology is designed to detect illegal content, the company acknowledged that it can sometimes generate false positives. A human moderator is reviewing the content, but an error occurs that causes the system to immediately ban affected accounts.

“We are working to provide better safeguards so this does not happen again,” the company wrote.

Our systems flag content by matching it to known harmful material. This type of similarity matching can lead to false positives, which is why a member of our Trust & Safety team always reviews flagged content before taking any action.

The intended behavior is…

— Discord Support (@discord_support) July 7, 2026

via X and Redditusers claimed that they were permanently suspended simply for uploading images containing square grid patterns. Many users have speculated that Discord’s AI moderation tools are becoming increasingly sensitive to grid-like patterns because they have previously been used in attempts to hide or hide NSFW and child exploitation content from automated detection systems.

Affected users have expressed their frustration on social media, with some arguing that a permanent account ban based solely on automated detection could have serious consequences, especially for users who rely on Discord for work, gaming communities, or long-distance social connections.

“Losing a Discord account over something unfair like this can be extremely devastating and hits users hard, and every day millions of users are affected by false AI bans. This needs to stop,” one X user books.

My account was wrongly banned from your platform due to a bug in your automated AI system that detected my game textures as CSAM. I need to recover my account because I am a game admin and use Discord for all my communications. You asked for my comment to be reviewed.@Disagreement @discord_support pic.twitter.com/QfAkCIJo6S

– Jdbryant 🎂 Today (@jdbryantdev) July 4, 2026

Discord isn’t the only one facing moderation issues due to automated systems. Last year, users Instagram and Facebook groups It reported widespread unjustified account suspensions that many believe were caused by AI moderation systems. Although users have pointed to automation as the likely culprit, Meta has not publicly confirmed whether AI errors are responsible. Now the oversight board is dead Pressure for increased transparency.

Tumblr last year too I faced complaints of users who said their accounts were suspended en masse without clear explanations.