Does the counter argument work?
IS ‘counter speech’ enough to tackle hate speech? This is what a senior Irish Twitter boss told TDs last week. If someone is being targeted by offensive tweets, a crowd rallying around the victim can provide restorative natural justice. The system rights itself.
It’s an American way of thinking (even though the executive who articulated it, Ronan Costello, is Irish). But is it right? The notion came up in a four-hour clash between TDs and the policy chiefs of Twitter, Facebook and Google.
The topic of Fiona Ryan and Jonathan Mathis, the mixed-race family who say they’re quitting Ireland over hateful comments and threats received after being in a supermarket ad, arose.
Fianna Fáil TD Jack Chambers asked whether merely deleting a hateful tweet, and not banning the Twitter user, was a sufficient sanction.
Twitter’s European director of public policy, Karen White, said it was one of a number of available sanctions. But then her Twitter public policy colleague, Costello, went further. He said that hate speech can sometimes be tackled by ‘counter speech’.
“When someone tweets something that the majority of Twitter users find distasteful or offensive, we often see a consensus of solidarity around the opposing point of view,” he said.
Asked by Chambers whether this wasn’t just outsourcing adjudication to a “mob”, Costello said that the principle of “counter speech” was something that is gaining acceptance.
“This notion of counter speech has been established for several years now,” he said.
Taken to its logical conclusion, this is a very American — rather than European — idea.
Speech is sacrosanct, the theory goes, so if someone screams equally loudly at an initial screamer, there is less of a problem due to the ‘balance’.
But should we really look on this as a positive thing because it brings out the solidarity in us with a fellow human under attack? Or just an algorithm for a shoutier, coarser, nastier public arena?
If someone racially abuses my friend, are 1,000 (or 10,000) retweets for a retaliatory tweet (which condemns the person who tweeted the racist remark) a mitigating factor for Twitter to take into account when deciding whether or not to merely delete the offending tweet or suspend the tweeter?
To be clear, there’s more to ‘counter speech’ than just equality of abuse. Recent European Council discussions have raised the idea of ‘counter speech’ as a broader initiative to tackle hate speech on a much more basic level.
I think this is what the Twitter executives meant also to highlight during their exchanges with TDs and Senators in the Joint Oireachtas Committee.
And it’s also fair to point out that the deeper discussion of the session, about how to make online networks less hateful and abusive places, is a difficult question.
It’s difficult mainly because few of us agree on what “hateful” or “abusive” really means.
For four hours, TDs and Senators asked the policy directors present to make things “better”. But when asked for guidance on specifics of what phrases or conversations to ban, the legislators usually come up short.
Nevertheless, deputy Chambers remained unhappy with the answers around the Ryan-Mathis case, saying that simple deletion of an abusive or racist tweet is “a really weak enforcement consequence for the person who’s brought such hatred to [victims’] lives”. He asked whether deleting a single tweet was “sufficient” in terms of enforcement or whether there should not be “a much greater consequence for the person who publishes that tweet”.
White again defended the social network’s level of enforcement.
“The reason that’s sufficient is that there is a purpose there in trying to educate that particular user that they have broken the rules,” she said.
“Consistent rule violations will result in permanent suspension. Then if law enforcement were to trigger an investigation on the back of behaviour like that, we can work with them as part of their investigation.”
When asked if self-regulation was working with regard to social networks, White answered that it “broadly” was. But while this might raise eyebrows, her answer has to be taken against the backdrop of a lack of definition provided from the elected officials grilling her.
One suggestion on a related theme was offered last week by the Garda Commissioner, Drew Harris.
The police chief said that a hate crime should be defined as one “perceived by the victim, or any other person, to be motivated by hostility or prejudice, based on actual or perceived age, disability, race, colour, nationality, ethnicity, religion, sexual orientation or gender”.
While this sounds like it leaves it up to the person affected whether it’s to be classed as a hate crime, it’s likely that it would still have to pass a further test (of sorts).
What of the famous artificial intelligence that is supposed to be making social networks a safer place?
They work much better for things like sexual abuse content than bullying or harassment, said Claire Rush, Facebook Ireland’s lead counsel for content and regulatory affairs.
“It’s much harder to train artificial intelligence or machine learning to tell if a hateful slur is being used in a derogatory way or in a comedic, satirical or self referential way,” she said.
“So that’s why the rates of AI detection are lower for content like bullying.”
This issue isn’t going to be solved soon.
Consensus is difficult to reach. Many reading this column believe, passionately and sincerely, that political or cultural opponents need to be banned from social networks for the good of society.
In some instances, they make a good case. But we can’t agree on which ones.