The European Commission announced on Thursday that social media platforms on the continent are removing twice as much “illegal hate speech” as they were six months ago. According to the EU’s new report, Facebook, Twitter, and Youtube have stepped up their efforts in ‘combating hate speech’.

Twitter was singled out for lagging behind the other players. While Facebook ad YouTube removed 66% and 66.5% of flagged ‘hate speech’ respectively, Twitter only removed 37.5%.

It cannot yet be proven, but it’s very likely that this week’s Twitter purge, which got Geert Wilders banned in Germany, Tommy Robinson banned altogether for 7 days and will probably result in the removal of Pat Condell’s account.

The publication of the report and Twitter stepping up its campaign, do seem to coincide.

Vĕra Jourová, EU Commissioner for Justice, Consumers and Gender Equality:

The results of our second evaluation of the Code of Conduct are encouraging. The companies are now removing twice as many cases of illegal hate speech and at a faster rate when compared to six months ago. This is an important step in the right direction and shows that a self-regulatory approach can work, if all actors do their part. At the same time, companies carry a great responsibility and need to make further progress to deliver on all the commitments. For me, it is also important that the IT companies provide better feed-back (sic) to those who notified cases of illegal hate speech content.

From the EU press release, we further learn that:

“1) On average, in 59% of the cases, the IT companies responded to notifications concerning illegal hate speech by removing the content. This is more than twice the level of 28% that was recorded six months earlier.

2) The number of notifications reviewed within 24 hours improved from 40% to 51% in the same six months period. (…)

3) As compared with the situation six months ago the IT companies have become better at treating notifications coming from citizens in the same way as those coming from organisations which use trusted reporters channels. Still, some differences persist and the overall removal rates remain lower when a “notification originates from the public.

4) Finally, the monitoring showed that while Facebook sends systematic feedback to users on how their notifications have been assessed, practices differed considerably among the IT companies. Quality of feedback motivating the decision is an area where further progress can be made.

The study concludes that:

“Xenophobia (17.8%), which includes anti-migrant hatred, has been reported, together with anti-Muslim hatred (17.7%), as the most recurrent ground of hate speech, followed by ethnic origin (15.8%).”