The researchers used an AI system trained to identify comments in the dataset that were likely to violate Meta’s current policies in three areas: violence and incitement, hateful conduct, or bullying and harassment.
Comments violating Meta’s policies regarding violent threats quadrupled, from 1,800 in the six months before the change to 7,600 in the six months. Hate speech comments also quadrupled, from 6,900 to 30,000. Comments breaking Meta’s rules on bullying and harassment doubled from 15,700 to 39,900.
“We regularly issue public reports tracking infringing content on our platforms, and the prevalence of hateful conduct did not increase throughout 2025,” a Meta spokesperson told WIRED. He said the company could not directly address the report’s claims without seeing the full research. WIRED provided a list of the derogatory comments cited in the report, but Meta did not comment on these. Hours before the report was published, several examples were removed from Facebook.
“When companies reduce oversight in areas like violence, hate, and harassment, it should be no surprise to see those harms grow,” Senator John Curtis, Republican of Utah and a member of the Commerce, Science and Transportation Committee, said in a statement to CCDH.
The data collected by CCDH researchers is reflected in Meta’s own transparency report from 2025, which shows how the company nearly halved its proactive content moderation enforcement in the months following its policy change. “The increase in abuse and the decline in enforcement track each other almost exactly,” the report’s authors write.
<a href