wired.com via Reddit

Meta Speech Rule Rollback Leads to Tripled Political Threats on Facebook — CCDH Study Finds Violent Threats Quadruple Against US Lawmakers in Six Months After Hate-Enforcement Cuts

meta ai ethics content-moderation platform-policy

Summary

A Center for Countering Digital Hate report documented that abusive and racist posts targeting US legislators tripled within six months of Meta removing automated hate-speech enforcement in January 2025, with violent threats specifically quadrupling; threats against President Trump doubled from 800 to 1,900 posts on Facebook in the same window. Meta replaced its AI-driven automated content moderation with a Community Notes model under chief global affairs officer Joel Kaplan's stated rationale that the company had been 'over-enforcing' and 'censoring too much trivial content.' The Wired investigation covers both Republican and Democrat lawmakers as targets of the surge in threatening content.

Shared on Bluesky by 5 AI experts