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

What happened – The Appeals Centre Europe, an independent dispute‑resolution body created under the EU’s Digital Services Act (DSA), released its annual report on moderation appeals for the period April 2025 – March 2026. The data show that the centre overturned platforms’ decisions not to remove reported hate speech in 70 % of the cases it could actually review. TikTok fared worst (83 % overturn rate), followed by Instagram (74 %), Facebook (61 %) and YouTube (58 %). At the same time, the centre received more than 14 000 suspension disputes but was able to examine the full content in fewer than 150 cases because the platforms failed to provide the requested material within the statutory 30‑day window.

Legal basis – The DSA obliges very large online platforms (VLOPs) to offer a “trusted flagger” system and an independent review mechanism for users who contest moderation decisions. Article 14 requires platforms to supply the contested material to the review body within 30 days, and Article 15 empowers the body to issue binding decisions that must be implemented without undue delay. Failure to comply can trigger administrative fines of up to 6 % of global turnover under the DSA’s enforcement regime. In addition, the General Data Protection Regulation (GDPR) imposes a duty of transparency on data controllers when processing personal data for content‑moderation purposes, meaning that refusing to hand over account‑related information may also breach Articles 12‑15 (right of access) and Article 5 (lawfulness, fairness, transparency).

Impact on users and companies

• Users – Victims of hate speech face a double‑layered barrier: first, the platform may leave the offending post online; second, the user’s appeal can be dismissed automatically if the platform does not supply the evidence. The centre’s report notes that 7 300 disputes were decided by default in favor of the user, simply because the platform missed the evidence deadline.
• Platforms – The overturn rates expose a systemic shortfall in policy enforcement. For TikTok, an 83 % overturn rate suggests that the majority of reported hateful content remained visible despite the company’s own community guidelines. Conversely, the fact that platforms also had decisions overturned when they removed content (52 % of reviewable cases) indicates inconsistent application of rules, which can expose them to claims of over‑censorship.
• Regulators – National DSA enforcement authorities in Ireland, Germany and France have already opened preliminary investigations into the “default decision” pattern. Under the DSA, regulators may issue interim measures and levy fines for non‑compliance with evidence‑handover obligations.

What changes are required

1. Full compliance with evidence‑handover deadlines – Platforms must build automated pipelines that retrieve the exact post, associated metadata and user‑account information within the 30‑day window. The DSA’s technical annex recommends using standardized APIs (e.g., the EU‑DSA Moderation API draft) to ensure interoperability.
2. Transparent reporting – Both the DSA and GDPR require clear, machine‑readable explanations for each moderation action. Companies should publish quarterly dashboards showing the number of appeals received, the proportion of default decisions, and the time taken to provide evidence.
3. Binding implementation of review outcomes – Article 15 of the DSA makes the Appeals Centre’s rulings enforceable. Platforms must develop internal escalation procedures that automatically apply the centre’s decision, whether it orders removal or reinstatement of content.
4. Enhanced oversight – Regulators may consider designating a “designated compliance officer” for each VLOP, tasked with monitoring DSA‑related KPIs and reporting directly to the European Commission’s Digital Services Coordinator.
5. User‑friendly appeal pathways – The centre’s free‑of‑charge service is a model, but platforms should embed a single‑click appeal button within the user interface, linking directly to the independent review body to reduce friction.

Broader context – The report arrives as the EU continues to push back against US‑based tech dominance. Recent actions, such as the European Commission’s investigation into Elon Musk’s X and the €120 million fine against the platform for advertising‑policy breaches, illustrate a growing willingness to enforce the DSA’s accountability provisions. At the same time, civil‑society groups warn that the “over‑censorship” narrative is incomplete; the data show that platforms are simultaneously under‑moderating hate speech and over‑moderating other content, a pattern that undermines trust in online discourse.

Next steps for stakeholders

• Consumers – If you encounter hateful content that remains online, use the platform’s built‑in reporting tool and follow up with a free appeal to the Appeals Centre Europe. Keep a copy of the original post (screenshot) to aid the review.
• Companies – Conduct an internal audit of your DSA compliance processes, focusing on evidence‑handover and default‑decision rates. Investing in a moderation‑as‑a‑service architecture that logs every moderation action in an immutable ledger can simplify evidence provision.
• Regulators – Continue monitoring default‑decision statistics and consider issuing targeted fines for systematic non‑compliance. The European Commission’s upcoming DSA Review (expected late 2026) may tighten the evidence‑handover deadline to 15 days.

The Appeals Centre Europe’s findings demonstrate that the EU’s moderation‑accountability regime is working, but only when platforms respect their legal obligations. Without full cooperation, the promise of an independent, user‑centric review process remains unfulfilled, leaving victims of hate speech without effective recourse.

---

Sources:

• Appeals Centre Europe 2025‑2026 Report
• EU Digital Services Act (DSA) – Official Text
• GDPR – Articles 12‑15
• European Commission press release on X investigation