This is Behind the Blog, where we share our behind-the-scenes thoughts about how a few of our top stories of the week came together. This week, we discuss getting back on our AI slop bullshit, deepfakes in schools, and epistemic virtues.
JASON: I was back on my bullshit this week, by which I mean staring at some horrible AI-generated shit on Facebook. In this case, I was looking at slop of ICE raids generated using OpenAI’s Sora. I talk about this in the article but one of the many reasons why I think social media—especially Facebook and Instagram—is fucked if they continue to monetize and promote this stuff is because this specific slop page is not being pushed by anyone who seems to have any sort of ideological interest in immigration or the United States or anything like that. It’s just someone trying to make money. And it’s almost definitely one single person, with one single account, generating millions and millions and millions of views. We are very early in the horrible AI slop game, and yet we are seeing the damage that just a few people can do with industrial-grade content generation machines.
Over the years I have talked to a lot of academics about the idea of “trolling” on the internet. The person who has most informed my thinking over the years is Whitney Phillips, who has written a lot about 4chan, “trolling” culture, and generally the worst corners of the internet in a series of books and papers. This Is Why We Can’t Have Nice Things, which is about 4chan, is extremely good, as is her paper called The Oxygen of Amplification. She (and other academics) have argued that it’s not the intent behind “trolls” that matters, it’s the impact. She wrote a lot about “just trolling” racism and misogyny and xenophobia on the internet and what’s important is how it impacts people and changes how they interact with online spaces. You can’t really be “ironically” racist, you’re just racist. And with anonymity and pseudonymity online, you can’t really infer what’s inside of someone’s heart because often it doesn’t matter if the person posting Harambe memes or making deepfake nudes of celebrities of their classmates or whatever didn’t actually “mean” it.