Théodore Cazals remembers the image that broke him.
Two South Asian kids, emaciated and obviously AI-generated, with full beards on little boy faces. One with no hands. One with one foot. Holding a sign that said "it's my birthday, please like." Sitting in the middle of a busy road in pouring rain with a birthday cake in front of them.
On Facebook it got nearly a million likes and hearts.
He was twenty. French. A student. And he did what a lot of us do when we see something stupid and infuriating online — he started an account to mock it. Called it "Insane AI Slop." Started calling out the fake, half-baked content that was fooling people by the millions. His inbox exploded. In a few months he had 133,000 followers.
I keep coming back to that story because it nails something I've been feeling for a while and couldn't name. We've entered a part of the internet where the content itself is the problem. Not the stuff we look for. The stuff that finds us whether we want it or not.
Zuckerberg calls it the "third phase" of social media. First was friends and family. Second was creators. Now it's AI, and he's all in on it. "Soon we'll see an explosion of new media formats that are more immersive and interactive, and only possible because of advances in AI," he told shareholders. YouTube's CEO said in December alone, over a million channels used their AI tools to make stuff. This isn't fringe anymore. This is the business model.
But business models work until they don't. And right now it feels like people are deciding they don't want this.
And the backlash isn't coming from critics or professors writing op-eds. It's coming from normal users who are just... tired.
Tired of scrolling past fake pictures of poor kids begging for likes. Tired of restaurant menus with AI photos that look like horror movie props. There are whole TikToks now dedicated to roasting "AI slop menus" as a sign a place is low-quality and lazy. By early 2026 Microsoft was pushing AI so hard online everyone started calling it "Microslop" — it got so common the name got banned in some of their official communities.
The numbers show it too. Gartner found 49% of US consumers now say generative AI has actively made content worse. Pew found people are getting more negative on AI and Big Tech in general. But under the surveys there's something more visceral. People are starting to treat AI slop the way we treat spam. Something to filter, ignore, mock. The novelty is gone. What's left is just noise.
I think this anti-slop revolt is bigger than just being annoyed at bad images. It's a reaction to something deeper. We've spent three years being told AI will make everything better — more efficient, more creative, more accessible. And in some ways, yeah, it has. But the actual experience of being online has gotten worse. The feeds are fuller, the images are faker, and the gap between what we see and what's real keeps getting wider. They promised abundance. We got sludge.
And the companies know it. That's why in July 2026 four of the biggest AI labs — Anthropic, OpenAI, xAI, and Google — signed that weird joint statement saying there's so much "enormous competitive pressure" that none of them can slow down on their own, so the government needs to step in and regulate everyone. They're basically saying: we can't stop ourselves. Please make us stop. And then they got sued for that too, with people arguing the agreement was collusion to give subscribers less value. So they're stuck. Public hates them, government might regulate them, shareholders want them to grow forever. Nobody's happy.
The most direct version of the pushback is QuitGPT. Since January, over 700,000 people have pledged to cancel. They claim they've knocked $1.5 billion off OpenAI's market cap. And it's not just about slop. It's about the military deals, the surveillance, the labor. One of their most viral posts was about Kenyan workers who have to sit and watch hours of explicit, abusive content to train the models — models that might one day be used to bomb their own countries. That's the dark side of the slop. Someone had to make it. Someone had to watch the worst stuff humans do so a chatbot can write you a birthday card.
So where does that leave us? I don't think AI is going away. I don't think the slop is going away either. But we are at an inflection point.
The people who make this stuff and the platforms that push it are betting that you don't care enough to push back. And right now, the pushback is coming from the edges — a twenty-year-old in Paris with an X account, anti-slop Facebook groups with hundreds of thousands of members, college kids and political campaigns who've started treating AI images as a liability, not an asset.
What strikes me most is how human the backlash feels. It's not organized. It's not funded. It's not really ideological. It's just people saying in a thousand different ways: I want to know what's real. I want to see something a person actually made and cared about. I want the mess, the imperfection, the proof that someone tried.
There's real AI polish fatigue now. The more perfect it looks, the less we trust it.
I don't have a solution. I don't think there is a clean one.
But the backlash is telling us something important. The problem with AI slop isn't the tech. It's the values behind it. The race for engagement. The need to fill every feed, every second, every scroll with something — anything — to keep us looking. Slop is what happens when you optimize for quantity over quality, volume over meaning, clicks over care.
Théodore's account is called "Insane AI Slop." He's making fun of it. But he's also archiving it. And in doing that, he's holding up a mirror to the platforms that profit from it.
And what you see in that mirror is a world where the content is fake, the engagement is hollow, and the only thing that's real is how exhausted everyone scrolling past it feels.
Eventually people stop scrolling.
That's the thing about slop. You can serve it by the ton.
You can't make anyone eat it forever.
Comments
Post a Comment