I’ve seen a lot of strange things covering tech and AI over the years. You get used to the weird, the wonderful, and the downright bizarre. But every now and then, something pops up that makes you stop and just say, "Wait, what?"
This is one of those times.
We've all been hearing about the incredible leap forward in AI video generation. Tools like OpenAI's Sora promise to turn a simple text prompt into a cinematic masterpiece. The potential is huge, right? For filmmakers, artists, and creators, it feels like a whole new world is opening up.
But here’s the thing about powerful new technology: it never stays in a sterile, controlled lab. It gets out into the wild, into the hands of everyone on the internet. And that’s where things are starting to get really, really weird—and frankly, pretty disturbing.
So, What’s Actually Going On?
On platforms like TikTok, a new and unsettling trend is bubbling up. People are using AI video generators to create content featuring AI-generated children. And we’re not talking about wholesome, animated stories.
We’re talking about videos that are intentionally shocking and deeply inappropriate.
Now, you might see the name "Sora 2" being thrown around. Let's clear that up. It doesn't seem to be an official sequel to OpenAI's Sora (which isn't even widely available to the public yet). It’s more likely a catch-all term people are using for a variety of accessible AI video tools that can create this kind of content. The specific tool is less important than what people are doing with it.
And what they're doing is creating a new, bizarre genre of shock content.
The Videos Are as Bad as They Sound
I’m going to be direct here, because trying to sugarcoat this doesn’t help. The content being created is designed to be as offensive as possible.
Think about the kind of ad you’d never, ever see on TV. Some of these videos are styled as fake commercials. One of the most-cited examples features AI-generated kids playing with adult sex toys as if they were regular toys. It’s jarring, it's gross, and it's clearly made to provoke a strong reaction.
It doesn’t stop there.
Other creators are making fake playsets themed around notorious figures. I’m talking about AI-generated "Jeffrey Epstein Island" or "Diddy's Fun House" toy sets, complete with smiling, synthetic children. It’s dark humor taken to a deeply uncomfortable extreme, using the innocence of children—even fake ones—as the punchline.
Why This Feels So Different and So Wrong
Look, the internet has always had a dark side. Shock content is nothing new. But this feels different. Why?
I think it comes down to a few things.
First, there’s the "uncanny valley" effect, but supercharged. These aren't just badly photoshopped images. The AI creates fluid, moving video of children who look almost real, but not quite. Their movements can be a little jerky, their faces a bit plastic. This slight "off-ness" combined with the horrifying subject matter creates a uniquely unsettling feeling. It’s like watching a nightmare come to life.
Second, it’s the sheer ease of creation. In the past, creating something like this would have required sophisticated CGI skills, actors, and a film set. It would have been a massive undertaking. Now, someone with a twisted idea can just type a few prompts into a tool and generate a video in minutes. The barrier to creating deeply disturbing content has basically been erased.
And that leads to the biggest concern: normalization. When you can create and share this stuff so easily, it starts to creep into the mainstream. It becomes a meme, a joke, a trend. And when you’re using the image of children for this kind of shock value, you’re crossing a line that we, as a society, have generally agreed shouldn't be crossed.
What Can Even Be Done About This?
This is the million-dollar question, isn't it? Platforms like TikTok are constantly playing a game of whack-a-mole with problematic content. They take one video down, and ten more pop up.
AI-generated content makes this infinitely harder. How does an algorithm reliably detect a fake child in a fake ad? How does it understand the nuanced, offensive context of a Diddy-themed playset? The creators of this content are smart; they'll find ways to tweak their prompts and visual styles to evade automated moderation.
This puts a huge amount of pressure on the companies building these AI tools. They have to build in better safeguards and restrictions. But that’s also a cat-and-mouse game. People will always look for workarounds, for the unfiltered model they can use to create whatever they want.
Ultimately, this is a conversation we need to have. We’re standing at the edge of a new era of content creation, and while it holds incredible promise, it also has a dark and ugly underbelly. We can't just be amazed by the technology; we also have to be critical of how it's being used. Because right now, it’s being used to create things that are just plain wrong, and we can’t afford to just look away.
