EU moderation watchdog says social media giants hate taking down hate speech

Personal Tech

Also somehow censoring too much while refusing to hand over account ban evidence for review

The EU's moderation appeals watchdog says Facebook, Instagram, TikTok, and YouTube routinely leave hate speech online while stonewalling attempts to independently review suspended accounts.

Appeals Centre Europe, an Ireland-certified dispute settlement body operating under the EU's Digital Services Act, says it overturned platforms' decisions not to remove reported hate speech 70 percent of the time between April 2025 and March 2026, including content targeting migrants, Roma communities, religious minorities, and LGBTQI+ people.

TikTok fared worst, with the Appeals Centre overturning its decisions in 83 percent of reviewed hate speech disputes, followed by Instagram at 74 percent, Facebook at 61 percent, and YouTube at 58 percent.

The report suggests that Europe's moderation accountability regime is functioning more or less how veteran tech policy watchers feared it might. In theory, the system gives EU users a free way to challenge moderation decisions through an independent review body. In practice, Appeals Centre Europe says platforms often fail to hand over the disputed content or account material needed to review cases.

Account bans appear to be where the process really starts to fall apart. Appeals Centre Europe says that despite receiving more than 14,000 suspension disputes, it got enough content from platforms to fully review fewer than 150 cases. Instead, thousands of cases ended in what the Appeals Centre calls "default decisions," in which platforms failed to provide the material within 30 days, and the ruling automatically went in the user's favor. More than 7,300 disputes ended that way.

Meta appears to be struggling most with the "independent review" part of independent review. The report says that out of more than 4,600 eligible Facebook and Instagram account suspension disputes, the company supplied the disputed content in fewer than 100 cases.

The findings also complicate the increasingly popular narrative that platforms are over-censoring everything in sight. Appeals Centre Europe says it overturned decisions to remove users' content 52 percent of the time in cases where it could actually review the material. Rules around restricted goods and services proved especially messy, with the group overturning removals in 65 percent of reviewed disputes.

In other words, the platforms appear capable of simultaneously leaving up content they should remove while deleting content they should leave alone, which is an impressive operational achievement even by social media standards.

Thomas Hughes, CEO of Appeals Centre Europe, said: "Online hate and harassment have real-world consequences for many people and communities. In more than two-thirds of our decisions about hate speech, we found that platforms failed to enforce their own policies and left up hateful content. This goes to show that platforms don't always get it right. If you're in the EU, you can challenge a platform's decisions free of charge to Appeals Centre Europe and get an expert, impartial review."

However, the report also suggests platforms do not always implement the Appeals Centre's decisions even after cases are reviewed. In one section covering disputes submitted by civil society groups, the organization says it is aware of only "a handful" of cases in which platforms acted on its rulings, while many disputed posts remained online.

That's not exactly the robust moderation accountability system Brussels has spent years promising users. ®