First Glance
I’ve written here before about how terrible my YouTube recommendations are. It’s a point of pride how little the algorithm seems to know about me. But things have gotten much worse in my feed, namely all the Kitchen Nightmares clips and Doctor Mike reaction videos have been replaced with nonstop coverage of how terrible and ridiculous Amber Heard is and Johnny Depp’s saintly affable behavior in putting up with her silly defamation attempt. I don’t actually think Heard is ridiculous nor do I find Depp affable, in fact I find his attempt to recharacterize Heard’s domestic abuse claims as “defamation” disgusting. But from looking at my YouTube recs you’d think I was mad as a hatter for the man, an absolute Edward ScissorStan. And according to several news sites, the “court of public opinion” agrees.
Double TakeThis term “court of public opinion” is strange to me. First of all, the source for most of these articles are “PR Experts” who, in a few cases, are speaking on the condition of anonymity which just makes me think they’re trying to conceal connections. But the bigger reason is that the data they’re using to poll this supposed court are hashtags and views of the same videos that are flooding my feed. In an instance where algorithms are serving trial content to victim-sympathetic folks like myself whose only social media interaction with Depp was 2006’s Jar of Dirt remix, are social media interactions an accurate gauge of public opinion?
Before I go on, get your scrolling finger ready because I want to show y’all the extent of content I’ve been served. These video recommendations are all from one visit to my YouTube homepage. I haven’t included the titles but the screengrabs speak for themselves:
Not only are they all very “Amber Heard bad woman Johnny Depp cool chill dude” but the occasional appropriative edits (see: acrylics & wig) are embarrassing.
The real question here is BOTS?! and the answer is yes-probz. The only folks that seem to have attempted a bot-calculation is a 3 person Israeli firm called Cyabra which found around 11% of total trial social media posts about the trial to be from bots. This represents 10% of pro-Heard activity, coming from “uninterested third parties looking to promote a product,” and 5% of pro-Depp activity which seems to be anything but uninterested given that 93% of all accounts analyzed were pro-Depp.
Taking Cyabra’s data at face value, 95% of the activity is from human people. This large proportion could indicate a genuine upwelling of support for Depp, but I don’t think that’s how viral internet content actually works. In fact, 5% of the 6.8 billion views that the #justiceforjohnnydepp hashtag has gotten on TikTok is still 340 million views, certainly enough to send a few videos on their way to pick up tons of human impressions. Cnet did some reporting on just how impactful this virality has been, interviewing two creators who pivoted their entire YouTube channels to capitalize on the virality of Depp content, one of whom has garnered so many views that he’s hoping to buy a home with the ad money. With a team of bots jump-starting your content to viral fame, it’s no wonder so many folks have flung themselves into the anti-Heard herd.
Hindsight
Social media algorithms have been in the hot seat several times over the past few years, with the 2016 election, Q-Anon, and several other disinformation campaigns and conspiracies. But beyond all those, this banal celebrity defamation trial shows just how easy it is to game the algorithm and how powerful it can be. I witnessed this firsthand the other day when I brought up my YouTube recs at the pub and it triggered an exhaustive anti-Amber Heard tirade from someone who could not have had less skin in the game. This trial is not important or even interesting. Depp is no OJ and celebrity defamation suits are a dime a dozen. And it’s not like there aren’t morepressingissues to be discussed and fought over. If the so-called court of public opinion is so easily swayed over such a trivial, settled matter, what happens when we’re brought into session for something that matters?