EU report: tech firms are now assessing 89% of flagged content within 24 hours and 72% of such illegal hate speech is removed, up from 40% and 28% in 2016
Context & Ripple Effects
This report closes a three-year loop the Commission opened when its first audit found Facebook, Google, Twitter, and Microsoft reviewed only 40% of flagged hate speech within 24 hours. Successive rounds of published monitoring moved the number to 59%, then to a 70% removal rate reported a year earlier, and the 2019 figures — 89% assessed within a day, 72% removed — mark the high point of that voluntary code-of-conduct arc.
First-order effects
- The signatory platforms gain breathing room: with the Commission's own data showing near-universal 24-hour assessment, the immediate case for legislating hate speech takedowns weakens.
Second-order effects
- The Commission extends the same playbook to harder cases, having already issued one-hour takedown guidelines for terrorist content, signaling that each improvement in the voluntary metrics becomes the baseline for the next demand.
Third-order effects
- The monitoring regime itself becomes structural — but fragile: a later EU report showed removals slipping back to 62.5% by 2021, evidence that gains resting on voluntary cooperation reverse without binding rules, strengthening the argument for the EU's shift to enforceable moderation law.
The trend: The EU's publish-the-numbers approach to platform content moderation is converting voluntary codes into the evidentiary groundwork for binding regulation.