digiQ’s 2025 Annual Hate Speech Monitoring report

In their words:

The findings point to significant compliance gaps and raise concerns about the practical effectiveness of reporting mechanisms for ordinary users. Our data show that hateful content flagged through specialised reporting channels for expert organisations is addressed highly effectively, with over 85% of illegal content removed. However, ordinary users experience long delays and, in many cases, no assessment of their notices. On Facebook, the median time for assessing illegal content reported via the Notice and Action mechanism (Art. 16 DSA) was 41 days. By contrast, while TikTok assessed notices very promptly (median time of 0.5 hours), nearly 80% of the content remained online - including cases involving incitement to violence and support for crimes against humanity.

Read the report (in English) here