Taylor Lorenz’s latest piece for the Washington Post warns that ‘the online incel movement is getting more extreme’, according to a new report from the Center for Countering Digital Hate, an American think tank, and its brand new Quant Lab.
The report is the ‘culmination’ of an investigation spanning 18 months, containing ‘more than one million posts’ from the web’s ‘most prominent [incel] forum’. The report was also covered by the BBC and the Times, despite the fact that these sorts of AI linguistic analyses of the very same forum have been done ad nauseum by researchers and students at various points in their academic careers.
But, according to a tweet from CCDH founder Imran Ahmed, the ‘new, never-seen-before [sic] data’ of this report is ‘absolutely vital’. So what did we learn? As is customary, the report opens with a content warning for its more sensitive readers before informing them that slurs and specific site names would be redacted, ‘in order to avoid promoting them.’ Next, the report presents some dramatic but hard-to-contextualise statistics, like that the word ‘kill’ was used 1,181 times in one particular month.
Context, in general, is conspicuously absent throughout; there’s no consideration of whether these posts about rape or violence might be ironic or absurd, as incels often are. Granted, an algorithm is incapable of detecting irony, but the human beings at the Quant Lab should have given it a shot.
‘Analysis of their discourse shows this core group poses a clear and present danger to women, other young men, and reveals an emerging threat to our children,’ the report adds. But how? Is there any correlation between all this nasty posting and real-world behaviour? There’s no data to support this bold assertion. In fact, a recent study into the incel community suggests otherwise. And so does the report itself: the main section includes a subheader titled ‘Links to Offline Violence’ that includes just one case of a man arrested for possession of an assault rifle and an ‘attempt to commit a hate crime.’ Apparently, members of the forum have speculated that this individual once briefly had an account on the site.
The fact is that, to date, not a single one of the perpetrators of an ‘incel attack’ has been linked to the website specifically. Both the report and the WaPo piece frequently cite cases in which law enforcement stated that perpetrators were not motivated by incel ideology, so then why do they claim that incels pose a “growing threat”? This conflation may be an oversight by Lorenz, but it seems a deliberate misrepresentation by the authors of the report, who are doubtlessly aware of such details.
Most galling, however, are the report’s final recommendations, which explicitly call on Big Tech to act: YouTube must ban all incel accounts; Twitter must shut down the main account for the site ‘named’ in the report; Google must de-rank incelosphere websites; and platforms must address ‘digital harms such as body image and mental health that can drive children into incelosphere communities’ — an awfully tall order, made without any suggestions for its implementation. And finally, of course, CloudFare ‘must stop providing services to the Incel Forum and feeder sites.’
Never mind the threadbare concept or the sloppy methodology, the only thing new about this report is how flagrantly it calls for outright censorship. It should be resisted, or better yet, ignored.