Discord Banned Around 8,200 Users for Posting Inoffensive Images of Grids
Discord said a bug in its automated safety system falsely flagged harmless images and is restoring accounts for more than 8,000 banned users.
- Discord acknowledged yesterday that a bug in its AI-powered moderation system mistakenly banned more than 8,000 users since May, incorrectly flagging harmless images as harmful content.
- The system failed to pause uploads, banning accounts instead of allowing human review from the Discord Trust and Safety team, letting the bug prevent automatic ban lifting.
- Users like JDBRYANT reported wrongful suspensions after GAME TEXTURES were flagged as CSAM, while Discord cofounder Stanislav Vishnevskiy confirmed at least 8,000 people posted "benign images" that were falsely flagged.
- Discord Support confirmed all affected accounts are currently being restored, stating, "We had a bug that caused the latter" when the system prevented bans from being lifted automatically.
- Similar mass-suspension issues struck Instagram, Facebook, and Tumblr last year, prompting the Meta Oversight Board to push for increased transparency in automated content moderation systems.
20 Articles
20 Articles
The platform acknowledged that its monitoring system was wrong with totally innocent content and left more than 8,000 accounts out of service.
Since May, over eight thousand Discord accounts have been suspended after uploading certain images.
TechCrunch
blkalerts.comDiscord admits AI moderation bug wrongfully banned users over harmless images
The company confirmed that the issue had been affecting accounts since May, with an additional 200 users banned over the weekend before its team identified and fixed the problem.
Discord Banned Around 8,200 Users for Posting Inoffensive Images of Grids
Like many online platforms, Discord has safety systems in place to automatically detect and flag abusive, illegal, and otherwise harmful content. This system bugged out in a very odd way over the past few months, leading to thousands of users being incorrectly banned. The situation provides a good chance to consider how online platforms automatically detect harmful images.
Coverage Details
Bias Distribution
- 63% of the sources lean Left
Factuality
To view factuality data please Upgrade to Premium
