EU says social networks remove 70% of hate speech, up from 59% in May, review 81% of reported content within 24 hours; Instagram and Google+ join the initiative
Katie Collins / CNET :
Context & Ripple Effects
The EU's hate-speech code began in May 2016 as a voluntary commitment by Facebook, Twitter, Google, and Microsoft (agreed with the European Commission), after a December 2016 review found only 40% of flagged material was assessed within 24 hours. A June 2017 follow-up showed removals climbing to 59% on average (the first year-over-year improvement).
This January 2018 review marks the code's expansion as well as its acceleration: Instagram and Google+ sign on alongside the original four, while the group collectively hits 70% removal and 81% same-day review. The trajectory holds through a 2019 peak before reversing — by 2021 the EU reported removal rates had slipped back to 62.5% (the annual moderation report).
First-order effects
- Instagram and Google+ are now subject to the same semi-annual EU reporting cadence as Facebook, Twitter, YouTube, and Microsoft, extending the code beyond its founding four companies.
- The original signatories can point to a third consecutive measured improvement — 28% to 59% to 70% removal — as evidence the voluntary framework is working ahead of any legislative push.
Second-order effects
- The Commission gains leverage: each improving scorecard strengthens its argument that voluntary codes deliver results, while each laggard platform (Google+ joining at the group's new baseline) faces public comparison against peers.
- Newer entrants inherit an audit regime built around the incumbents' moderation infrastructure, pressuring them to scale review capacity quickly or appear as the code's weak link in the next report.
Third-order effects
- If the pattern holds, voluntary scorecards become the de facto regulatory instrument — the EU measures compliance publicly, platforms compete on the numbers, and formal legislation codifies what the metrics already enforce. The 2021 decline suggests the mechanism depends on sustained political attention rather than self-enforcing momentum.
The trend: Platform content moderation is shifting from ad hoc corporate policy toward government-audited, metric-driven compliance regimes, with the EU's semi-annual hate-speech reviews serving as the template.