,
32 tweets,
5 min read
Read on Twitter
A few thoughts (THREAD INCOMING) on the Maza/Crowder/YouTube debate: I think this ends up being tricky for companies in large part because they aspire to uniform rules, but the harms of abusive speech in a social network ecology vary wildly depending on the speaker’s audience.
Maza’s complaint isn’t fundamentally that some asshole directed some homophobic slurs his way. If a Twitter egg or rando blogger had used similar langugage in the course of criticizing his videos, I assume he’d say “what a jerk” & forget about it.
Rather, he’s (very understandably) angry that some percentage of Crowder’s millions of followers are then inspired to bombard Maza with hateful messages—any one of which he might shrug off, but which in the aggregate make his online experience intolerable.
I don’t know YouTube’s internal process, and I’m not interested in defending their decision here. Maybe they just said “We like money and Crowder drives traffic.” But I can imagine a somewhat more understandable rationale for their hesitation...
To wit: “Do we want to commit to a universal standard where where we’ll pull down a whole video making a political argument if there’s an offensive joke about someone’s voice, regardless of whether it causes harassment or harm?” Or alternatively...
“Do we want to start factoring in conduct that happens off our platform in deciding whether content violates our standards? And if so, what’s our standard for what kind of off-platform harms are bad enough, and how to determine a particular video or channel is responsible?"
Twenty years ago, this was less of a problem, because speakers who reached audiences in the millions were on very different platforms with different rules. ABC or the NYT weren’t going to let you call anyone a “lispy queer.” Even Fox might not invite you back.
And, of course, whatever the size of your audience, it was in practice a lot more difficult for the audience to decide to pile on the target of that sort of mockery—certainly not by the hundreds or thousands.
Intuitively, what we consider unacceptable harassment vs legitimate critique has a lot to do with power and reach. If Sean Hannity makes a habit of picking on some specific student activist, that’s abusive & irresponsible. If the student makes videos attacking Hannity...
…for an audience of a few hundred or a few thousand people, that’s legitimate criticism. Even if they’re using more or less equivalent language.
But recognizing the importance of power differentials, even apart from the practical difficulty of measuring them, very much goes against the DNA of social media platforms.
Remember the appeal, back when we were mostly a lot more sanguine about the promise of the Internet to connect humanity, was erasing traditional speaker/audience distinctions. We’re all just users here. Today’s egg may be tomorrow’s blue checkmark.
And that’s led most of the major platforms to be pretty strongly committed to a uniform set of standards focused on the content itself, without regard—at least officially—to the size of the speaker’s mic or the practical difference this makes in external consequences.
And maybe that just doesn’t make a lot of sense anymore, especially given the relative ease of distributed harassment. Maybe the standard appropriate to someone venting on their webcam for an audience of a few hundred permits too much harm scaled to millions of viewers...
…and the standard appropriate to someone broadcasting to a massive audience is either too restrictive or too unwieldy to apply to every user spouting their thoughts into a laptop.
And I say “maybe” because I’m not at all certain whether introducing some kind of tiered standard indexed to audience size WOULD be a better approach. But I don’t get the sense it’s even on the platforms’ radar, at least as a formal policy, even if de facto they do factor it in.
I should add that factoring in the status of the *target* makes this even trickier. Carlos Maza is doubtless what the law would call a “public figure.” But I’m a lot more sympathetic to his concerns than I would be if the target were, say, a Senator or a major TV personality.
Given Devin Nunes’ lawsuit-happy approach to online criticism, I’d be skittish about a standard that treats a complaint about “harassment” from someone with real power the same way it does a complaint from a moderately famous-for-DC journalist.
But once you go beyond a mechanical “audience size” approach and try to gauge relative “power” in some broader sense, it starts looking infeasible for platforms to make those sort of judgments at scale.
Well, speak of the devil. Looks like YouTube has decided to split the baby:
In a way, this is a form of distinguishing between speakers by audience size: Demonetization or denial of access to a “partner program” are only meaningful penalties for someone drawing enough traffic to make significant revenue from videos.
Though, predictably, accurately identifying harmful content is not alway easy:
The details of the Maza/Crowder case may actually make it *relatively* easier, because YouTube can just go with a mechanical rule against homophobic slurs. I’m not sure that gets at the heart of the real problem though.
Imagine Earth-2 Crowder who routinely attacks Maza—maybe mockingly imitates his voice—but never *explicitly* references his race or sexuality. The literal text is strictly about his work or ideas. And suppose similar distributed harassment still occurs.
(Maybe that’s a bad assumption, and the harassment is heavily contingent on the explicit homophobia, but it seems plausible it might still be pretty bad.)
From Maza’s POV, it might not make all that much difference to the experienced harm—the mob is still almost as ugly. But from YouTube’s perspective, it seems like a much stickier call without the bright line.
Intuitively a good compromise solution here is “set a relatively high bar for pulling content or banning people, but establish stricter standards for what gets actively promoted.” But that has its own set of difficulties.
You still have a scale problem—is the site ONLY going to surface manually curated content? If so, it gets a lot harder to break through, especially for particularly timely videos. If not, you need to hand-tweak the algorithms, which is inherently opaque...
…at which point folks start muttering about “shadow bans.”
Also, stating the obvious, but whatever you think of the merits of YouTube’s reversal here, it’s getting progressively harder for the platforms to sell the official line that they’re just impartially applying neutral community standards.
A very last point: For the folks who think “regulate the platforms like public utilities” is a good idea, recognize that it would moot this entire conversation about the parameters YouTube should establish for acceptable speech.
As soon as YouTube can be regarded as a state actor, the question is settled: Hate speech much, much worse than Crowder’s is entitled to full First Amendment protection, and measures to censor or even disfavor it get struck down easily.
