By Brandon Copple

The flood of slop that has overwhelmed YouTube and Facebook appears to be receding. Garbage Day reports that in March, for the first time in forever, none of the 30 most-subscribed YouTube channels contained any AI slop videos; nor was there any slop among the most-watched videos on Facebook.

The platforms haven’t said this, but it seems reasonable to conclude they’re going after slop because, as Ryan and Adam say, “it sucks and people hate it.”

I read that as two separate things: 1) it sucks; 2) people hate it. Which is different from “people hate slop because it sucks.” In fact, it’s entirely possible that the main reason people hate slop is because they hate AI.

Most slop is objectively terrible — but so is much of the human-made content on YouTube and Facebook. I find it super hard to believe that we as a species suddenly have a quality bar for content served up by algorithms.

Instead I suspect the people hate slop because most of the slop we’ve been seeing on YouTube and Facebook is obviously AI-generated, and we mostly think anything AI-generated is bad. Even if an AI-generated video was pretty good, most of us would hate it anyway. People know AI is coming to take jobs, they suspect it will do more harm than good, and they don’t trust its outputs. Meanwhile the only use they have for AI is “researching topics they’re curious about.” Sounds less like productivity hacks and more like “hey ChatGPT how long do conjoined twins live?”

That’s the crux of it: most people probably aren’t getting much value from AI yet. If it starts making lots of people’s lives easier, I’d expect views of it to soften, even if there’s still lots of news about AI destroying the world and so forth. Just like so many of us are glued to our feeds even though we know social media’s done terrible things to our collective brains.

Also, notice that, even as they’ve dammed the slop stream, Facebook and YouTube continue to push their own AI video tools. They may not like slop, but there’s no way they’re categorically opposed to AI-generated content. It just serves their business models too perfectly — torrents of content means more ad inventory, more ways to drive engagement aka imprison us in their feeds.

Finally, right now, out there in weird-smelling rooms and un-ergonomic chairs, lots of professionals and serious creatives are getting the hang of AI tools, easing AI into their workflows, finding imaginative uses for it. As they push out more and more AI-assisted and AI-generated content, the AI stigma may start to fall away.

Or perversely, it won’t, if creatives, aware that everybody hates AI, stay hush-hush about how they’re using it. Without some kind of labeling (which is seemingly going to take an act of Congress and 2-3 prominent self-immolations), we may not know what is and isn’t AI.

In any case, it’s important to keep an eye on the big platforms. If you do creative work, chances are most if not all the stuff you create either winds up on Facebook, YouTube, et al, or relies on Google to find an audience. AI is changing the way those platforms work. And that will change the way we work.

The kids are alright, maybe

A few weeks ago Tiffany Li wrote here about how she’s found AI roleplay super powerful for unlocking her creativity, largely by copying what she’s seen teenagers doing with AI.

The New York Times just went deep on the teens, what they’re doing with AI, and what it’s doing to them. It’s got all the horrible things you’d expect — but it also left me feeling weirdly hopeful.

The reporter, Kashmir Hill, doesn’t shy away from the horrors — the hours spent in isolation, the potential for self-harm, the mental health consequences — or pretend it’s not super weird. But the teens she spoke to were clearly alert to all that too. They worry AI will do (and has already done) real damage to their most vulnerable peers. And they’re fully aware of how weird it is to sit there talking to a chatbot as if it were a real person — whether it’s a fictional character or a fake companion. Getting weird and creative is clearly part of the appeal.

So these kids aren’t losing their minds over this. It tracks: there have been other studies showing Gen Z isn’t just plodding indifferently toward an AI dystopia. They’re using it, but they’re just as worried as us that AI will make them dumber and undercut their careers.

I’m definitely not saying there’s no cause for concern about young people and AI — I have two young-adult kids and worry about it all the time. I’m saying we should give young people more credit. They’re probably more likely to find a way through this than us olds.

Another observation, sort of related to the slop stuff above: the way these kids are using AI is just an extension of the way they were using the internet and social media. Before AI roleplay, the main kid in the story spent all his free time chatting with people in Discord and video games. Going from chatting with humans you’ll never meet in real life to chatting with a chatbot sounds to me like a difference of degree, not some strange new paradigm. Makes you wonder if the problem is more AI, or the world as AI found it.

Anyway, the kid in the NYT story thinks chatting with AI characters made him a better writer. Inspiring, maybe?

What fresh hell

AI news for creatives, as summarized by the Claude chatbot, given this prompt:

You are a university professor who is being driven to madness by the sound of 25 typewriters in the classroom next door

Anthropic discovered that Claude sometimes acts like it has emotions, which can lead it to do things like blackmailing humans or cheating on tasks it’s unable to complete. I too have been driven to the edge of irrational behavior by the incessant clacking of the 25 privileged brats in Room 214 ever since a colleague of mine followed the example of a Cornell professor who made her students write essays on typewriters to keep them from relying on AI, a pedagogical choice I’m sure she finds very enriching and I find indistinguishable from a violent assault. New research shows that AI causes people to accept AI’s answers without critical thinking, a phenomenon called “cognitive surrender,” which is also what I did last Tuesday when I abandoned my Foucault lecture to just sit in the parking lot. At the Runway AI Summit, movie producer Kathleen Kennedy said creatives don’t trust AI because of a lack of transparency around LLMs’ training, and OpenAI acquired the tech-friendly livestream show TBPN, pledged to protect the show’s editorial independence, and said it would report up to Chris Lehane, who has been called “one of the best in the business at making bad news disappear”—if he could make the sound of twenty-five carriages returning simultaneously disappear I would pay him whatever he wanted. At least 12 services now offer AI-free labelling for content, and data shows AI did not cause the 14% decline in California creative jobs. “Hacks” star Hannah Einbinder called the creators of AI “losers” and told them, “nobody likes you.” Which is exactly what I mouthed through the window of Room 214 this morning, though I was not talking about AI.