An overly heavy-handed, proscriptive approach from legislators here would undoubtedly promote a backlash. It is hard to imagine much consensus on an obscure European committee deciding what is and isn't acceptable, except perhaps at the very extremes.
However, notes Williams, Germany has managed it - and France is considering it - while the Online Harms white paper in Britain laid out some groundwork towards prevention in this space.
"Watch this space," he says. "If things carry on going the way they're going, then I wouldn't be surprised if the EU legislated in some way. The EU was one of the first to lead a convention on cyber crime, and then the 2004 protocol on xenophobic speech came out as an addendum to the cyber crime convention.
"I think it was very clear in terms of what the EU's intentions were for regulating, insisting that its member states regulate along the lines of protocol. The UK hasn't done it properly, there's lots of states that still haven't done it properly, but about half of them have integrated the additional protocol on xenophobic communications. A lot of it is about Jewish identity and holocaust denial, but also some of it is about race.
"So they were some of the first to actually make moves like this, and I wouldn't be surprised if there were additional moves in the years ahead."
Such efforts would be a swing back towards favouring regulators in general, an environment that the social media companies were not established in and did not flourish in.
"There wasn't any kind of national or supra-national organisation that could impose regulations on them," says Williams. "I think that's changing now."
This story, "Cardiff University's HateLab unearths online hate speech using AI" was originally published by Techworld.com.