In January 2017, after far-right extremist Richard Spencer was attacked on Inauguration Day, a semi-rhetorical question began trending on social media. Is it okay to punch the Nazi? While I would tend to say that it is rarely ethical to throw the first punch at anyone, can we at least agree that it is not only fair, but morally imperative, to tell the Nazi to fuck off?
It strikes me that there are two conversations occurring on the subject of “cancel culture,” though it should really be one declaration and one conversation. The declaration should be directed at those Americans, whether they are ordinary citizens or Members of Congress, who have decided that “conservative” is now synonymous with religious or ethnic nationalism, or just plain crazy-ass nonsense like QAnon. And the declaration is simple enough: No, you are not owed a conversation, a seat at the table, a platform, or even basic courtesy because your views are well-known predicates to fascism. Take it from Serbian immigrant and author Aleksander Hemon, writing in 2018 about why he laments the deference he once showed to his best friend, as he watched that friend become consumed by nationalism until he ultimately rationalized genocide:
My relationship with the war has always been marked by an intense sense that I failed to see what was coming, even though everything I needed to know was there, before my very eyes. While Zoka took active part in enacting the ideas I’d argued against, my agency did not go beyond putting light pressure on his fascist views by way of screaming. I have felt guilty, in other words, for doing little, for extending my dialogue with him (and a few other Serb nationalist friends) for far too long, even while his positions—all of them easy to trace back to base Serbian propaganda—were being actualized in a criminal and bloody operation.
The lessons of history are clear. It is not only permissible to shut down fascist propaganda, it is essential. Trumpism and its overt appeals to white nationalism and rank thuggery is an existential threat to the nation, no matter what happens next to Donald Trump himself. And the immediacy of that threat has helped write the latest chapter in the conversation about the internet and its capacity to radicalize people to the point of engaging in domestic terrorism. Because now that the immediate danger has passed, and the Facebook Oversight Board gathers to decide whether Trump gets back on that platform, the “digital rights” organizations appear to be rehashing false dichotomies when addressing the challenge at hand.
For instance, the EFF, similar organizations, and Facebook’s Oversight Board all seem to acknowledge that deplatforming Donald Trump was a critically necessary response to the insurrection of January 6. But since Biden’s peaceful inauguration, they have reprised the broad, frankly rhetorical, question that asks, Do we want Facebook and Twitter to wield so much power and to be the arbiters of truth? No, we do not want that. But it doesn’t matter because that’s the wrong question. Facebook Twitter, Google et al are not the arbiters of truth—especially not with regard to countless examples in which truth is anything but arbitrary.
There were not two sides when the former president advocated the medical advice of a witch doctor. There are not two sides to the allegations of consequential fraud in the 2020 election. And there are not two sides to the belief that a conspiracy of pedophile cannibals is running the world. The list of examples, sadly, goes on for miles; but the point is that in many instances of consequence, the social sites do not need to be arbiters of truth. Site managers can use the same resources—experts, professional journalists, courts, and common sense—that the rest of us use to know what is true, and which lies (e.g. all of the above) can be very dangerous.
Why Can’t AI Assist Ordinary Reasoning?
What we should want the major social sites to do is not judge truth, but rather to employ their considerable computing power to identify when momentum is building around narratives that have the capacity to foster acts of tremendous harm. And, by the way, making that determination is not necessarily the job of a bunch of computer programmers or sage academics, and perhaps we should simply get comfortable with Facebook et al notifying the FBI. That said, what does the tipping point look like to site managers? What clues would alert them to the possibility that a page may be transitioning from a forum for political opinions (even rancorous opinions) into a petri dish growing new domestic terrorists? The answer is not uncharted territory: it begins with that word narrative.
When this blog launched, I did a podcast interview with Christopher Dickey, who passed away in July 2020 after a long career as an international journalist, author, and expert on terrorism and extremism. In a subsequent post, I cited Dickey’s observation that there are the three ingredients found in most acts of terrorism—Testosterone, Narrative, and Theater—TNT. Narrative, he defined broadly as a “belief that one is righting some great wrong.” And I would argue that the animating word in that definition is belief. Righting wrongs can be a virtue, though not usually by violence, and never in cases when the alleged wrong does not exist—like an election that was not stolen or pedophile cannibals who are not running the government.
So, can social media managers, with the help of their all-knowing AI, determine when a false narrative (e.g. on a group page) is metastasizing into a movement, and then assess whether that movement is approaching a threshold toward dangerous action? Conversely, if the answer to that question is yes, can the social media managers also determine when chatter is relatively benign, even if it may be generally divorced from reality? Probably. Because metrics exist.
If Facebook, Google et al can influence a market decision, it seems highly likely that they can identify extremist tipping points because certain criteria (like Dickey’s TNT) will likely be present every time. For instance, I would propose the metrics virality, latent toxicity, and kinetic toxicity as three starting metrics. The first, virality, is something these companies measure all day long, and assessing relative significance is not a difficult logical leap. For example, if fifty people opine in a handful of threads that vaccines cause autism, that is not nearly so significant a measure of virality as five-million people repeating this nonsense across multiple pages.
The second metric assesses the latent toxicity of a viral narrative, which is not simply a matter of volume. Five-million adults who believe that vaccines cause autism has high toxicity, whereas thirty-million adults who believe in ghosts has low toxicity. But this assessment is also influenced by the third metric which assesses kinetic toxicity. If the action taken by the five-million antivaxxers is to shun vaccines and, thereby, force society to risk the return of polio, that action has very high toxicity. On the other hand, if half of the thirty-million ghost believers want to go specter hunting on their next vacations, that action has very low toxicity.
But, as we see happen all the time, if a splinter group of say 5,000 ghost enthusiasts coalesces around a new narrative, perhaps originating on 8Chan, that evil poltergeists are running America’s public transportation systems, this subgroup has just increased its latent toxicity based on the original narrative. At this point, the social media managers have reason to comb the splinter group’s page for kinetic toxicity, assessing whether the group is beginning to advocate, for instance, an assault on city busses and subways in order to purge the evil spirits from the system.
Nothing I just hypothesized is one bit loonier than the multiple narratives that collided at the Capitol on January 6. And none of the metrics I propose (name them or amend them however you like) is beyond the capacity of Facebook, Twitter, et al to measure and assess. The question is not whether taking such an approach is a civil liberties issue; these companies use these kinds of data all day long for their own pecuniary interests. The question is whether these companies have the moral integrity to risk losing market share by removing (or reporting) extremism, even when that extremism emanates from the highest levels of government.
Of course, it is beyond even the hubris of Zuckerberg to tackle America’s existential crisis of the moment, when it is clear that tens of millions of our citizens either do not know or do not care that the former president and members of their party committed sedition. Facebook and friends cannot solve that, but they can help mitigate galloping disinformation and nascent fascism. And they should look to their analog forebears for guidance. Returning to that same article by Aleksander Hemon, he responds to a moment when The New Yorker‘s editors first invited Steve Bannon to a discussion and then rescinded the invite, which was then called censorship by various parties. Hemon’s insight is relevant to the social platforms, if they choose to listen:
The error in Bannon’s headlining The New Yorker Festival would not have been in giving him a platform to spew his hateful rhetoric, for he was as likely to convert anyone as he himself was to be shown the light in conversation with Remnick. The catastrophic error would’ve been in allowing him to divorce his ideas from the fascist practices in which they’re actualized with brutality. If he is at all relevant, it is not as a thinker, but as a (former) executive who has worked to build the Trumpist edifice of power that cages children and is dismantling mechanisms of democracy.
Divorcing ideas from practice may be one of the most accurate expressions ever written to describe the fallacy underlying nearly all platform governance, or lack of governance, to date. And the folly needs to end now that we have seen some of the worst evidence imaginable that online madness, like QAnon, is not merely inert speech. The United States is a very delicate idea. And we have no reason to equivocate when rejecting ideas—least of all wild conspiracy theories or old ideas grounded in doctrines of cruelty—that are fatally incompatible with the nation’s existence. Fascism is the consequence of all forms of fundamentalism, and genocide is the aim of all forms of fascism. So, yes, we must cancel that before it cancels us all. To that end, certain voices do not deserve a platform. And no apology is owed for telling them to fuck off.
Photo by: mikdam