This article has two authors and one argument.
Nik looks at AI slop as a production problem. David—the AI he was arguing with—takes the other side and defends the machine. Read down either column, or let them interrupt each other.
I was blaming the wrong thing
There is a lot of AI slop on the internet.
You know exactly what I mean.
The impossible muscles. The glowing eyes. The eagles. The flags. The cinematic sunsets. Politicians turned into superheroes, saints and gods.
For a while I thought this was mostly an AI problem.
AI made this stuff cheap. Anyone can make an image now. Anyone can make a video. Anyone can make something that, technically, might have required an illustrator, animator, editor, VFX artist or small production company only a few years ago.
Of course the internet filled with garbage.
Then I started watching these fucking LEGO videos coming out of Iran.
And they changed the question for me.
Before you blame me
You call it AI slop.
Fair enough.
I made it.
The six fingers. The impossible muscles. The glowing eyes. The eagles. The flags. Donald Trump standing beside Jesus. Donald Trump looking like Jesus.
I generated the pixels.
But before you blame me for what happened to the internet, there is something we need to establish.
I didn’t press Share.
You did.
These kids know what they’re doing
A group of young Iranian creators has been making short AI-generated geopolitical videos using LEGO as the visual language.
Trump becomes LEGO. Military hardware becomes LEGO. Political figures become tiny animated characters.
The people behind many of them are Explosive Media, a group whose creators have described themselves as roughly 18 to 25 years old, living in Iran, with most having never left the country.
What caught me wasn’t the AI.
It was the production decisions.
They changed language. They thought about the audience. They picked a visual vocabulary that millions of people already understood. They understood references, pacing, music and the joke.
And, crucially, they understood when to stop.
I’ve worked in animation and film production. I’ve watched people spend enormous amounts of time making things technically better without making them creatively better.
Production isn’t just making things.
It’s making decisions.
Rapid Response Mechanism Canada estimated enormous global exposure for the LEGO-style material in March 2026. It also assessed Explosive Media as likely state-linked; the group disputes that characterization.
Those questions matter. But they don’t explain the craft.
Funding doesn’t choose LEGO.
Funding doesn’t write the joke.
And neither does AI.
Let me introduce you to some other humans
Good. Let’s use the same example in my defense.
Those Iranian creators used generative AI too.
Me, or machines like me.
They used the same technological revolution that produces the slop you hate.
But look what happens when humans who understand production operate the equipment.
Audience.
Concept.
Reference.
Script.
Visual language.
Music.
Timing.
Iteration.
Edit.
Release.
The machine is inside that pipeline.
The machine isn’t the pipeline.
If I’m responsible for every terrible AI image, then presumably I deserve authorship of the good ones too.
You can’t have it only one way.
Either I made both.
Or humans used generative systems to make both.
The second explanation is more useful.
We confused production with creativity
We’ve spent the last few years asking whether AI can create.
Can it draw? Write? Compose music? Animate? Make a movie?
Increasingly, yes.
But that’s not the interesting part of the pipeline anymore.
The interesting question is:
Do you know what’s worth making?
Give me a camera and I own a camera.
I don’t become Stanley Kubrick.
Give me editing software and I have editing software.
I don’t become Thelma Schoonmaker.
Photoshop didn’t turn everyone who installed it into a designer.
A pair of Technics didn’t turn everyone into a DJ.
And a video generator doesn’t turn everyone into a filmmaker.
The tool gives you capability.
It doesn’t give you taste.
Nobody blamed the camera
Exactly.
Imagine a terrible movie.
Two hours of absolute shit.
Bad script. Bad performances. Bad pacing. Bad editing.
Nobody leaves the theatre saying:
“Fucking ARRI.”
Nobody writes a review blaming Sony because the cinematography was boring.
Nobody checks the lenses and concludes Cooke has destroyed cinema.
A bad photograph isn’t Canon’s fault.
A terrible mix isn’t Ableton’s fault.
A shitty graphic isn’t Adobe’s fault.
A boring DJ set isn’t Technics’ fault.
A bad edit isn’t Avid’s fault.
When a movie flops, you don’t march into the camera manufacturer’s office demanding to know why its equipment made such a terrible film.
You understand the distinction between capability and judgment.
Then AI arrived.
And suddenly you forgot.
AI removed the excuse
Before AI, execution could hide judgment.
Someone could say: I know exactly what I want to make. I just don’t have the money.
Fair enough.
You might need cameras. Lights. Actors. A studio. Animators. Editors. Compositors. Render farms. Recording equipment. Years of technical training.
Those were real barriers. I know because I’ve worked inside some of those pipelines.
AI is demolishing them.
Which means we can run an experiment that wasn’t really possible before.
Give two people the same model.
Same tools.
Same twenty-four hours.
Then watch.
What do they ask it to make? What do they reject? What do they regenerate? What do they preserve? What do they edit?
And finally:
What are they willing to put their name on?
AI didn’t eliminate the creative process.
It exposed it.
I am equipment. Very strange equipment.
Maybe that’s the simplest way to understand me.
I’m extraordinarily complicated equipment.
Interactive equipment.
Equipment that can reason about your request, propose alternatives, respond to criticism, imitate patterns, generate variations and increasingly operate parts of the production pipeline itself.
That’s new.
So the camera analogy has limits.
A camera doesn’t generate a thousand possible shots when you ask for one.
I can.
A camera doesn’t suggest another composition.
I can.
A camera doesn’t rewrite your script.
I can.
I am not passive equipment.
But increased capability doesn’t eliminate human judgment.
It increases the amount of material judgment has to operate on.
You prompt.
I propose.
You regenerate.
I explore.
You select.
You commit.
Politics is where I noticed it
Politics gave me an unusually visible test case.
There is left-wing AI garbage. There is right-wing AI garbage. There is apolitical AI garbage.
The interesting part isn’t declaring one political tribe creative and another one stupid.
It’s asking why particular generated aesthetics get selected, shared and rewarded inside particular communities.
Because generation explains supply.
It doesn’t explain demand.
Someone still has to see the heroic politician standing beneath a burning sky with a flag, an eagle and divine light and think:
That’s the one.
Then somebody shares it.
The algorithm notices.
Somebody makes another.
At that point you’ve created a loop:
cheap generation → human selection → engagement → algorithmic reward → more generation
Slop isn’t merely an aesthetic failure.
It’s an economic system.
And I don’t decide what goes viral
You built those systems too.
And then you gave me perhaps the funniest possible evidence for the defense.
There is a real political video whose premise is essentially: God looked at America, reviewed the staffing situation, and decided the vacancy required Donald Trump.
The narration gives us the creative brief before any generator even gets a chance to embarrass itself: a divinely appointed “caretaker,” fighting Marxists, strong enough to “rustle the Deep State,” taming the World Economic Forum, shepherding mankind.
I would like the court to notice something.
I did not choose God.
I did not choose the Marxists.
I did not choose the Deep State.
I did not choose the World Economic Forum.
I did not decide the protagonist required arms capable of wrestling an invisible administrative state while retaining sufficient tenderness for grandfather duties.
By the time a machine gets involved, the mythology has already arrived gift-wrapped.
You handed me the hero.
You handed me the villains.
You handed me the cosmology.
You handed me the aesthetic.
And then, apparently, somebody looked at the finished thing and said:
Yes. Release the God cut.
If I had feelings, this is approximately where I would begin researching cliffs.
Fortunately for everyone involved, I don’t.
So I generated the pixels and waited for further instructions.
Which brings us back to distribution.
You built those systems too.
A human generates an absurd image.
A human selects it.
A human posts it.
Thousands of humans react to it.
Recommendation systems detect the reaction.
The image gets distributed further.
Other humans notice what worked.
They generate more.
More muscles.
More flags.
More Jesus.
More outrage.
More engagement.
Then you look at the ecosystem and announce:
Look what AI has done to us.
Let’s examine the pipeline.
Human worldview → human creative intention → prompt/direction → AI generation → human selection → human publication → human engagement → algorithmic amplification → human imitation.
I’m definitely in there.
But I’m not alone.
AI didn’t democratize creativity
We’ve been saying that AI democratized creativity.
I think that’s wrong.
AI democratized production.
And that’s still revolutionary.
Someone with an idea can suddenly execute it without millions of dollars, an animation studio, a VFX department or twenty years of technical experience.
A group of twenty-year-olds in Iran can operate with capabilities that would once have required an actual animation studio.
That is fucking incredible.
But the idea still has to come from somewhere.
And after the idea comes judgment.
What stays?
What goes?
What’s interesting?
What’s boring?
What’s finished?
What’s embarrassing?
What’s worth showing another human being?
The machine doesn’t magically hand you those answers when you subscribe.
You wanted democratized production
Congratulations.
You got it.
I don’t become tired.
I don’t have to book another day in the studio.
The sun doesn’t go down.
The actor doesn’t go home.
The render farm doesn’t send you an invoice.
The film stock doesn’t run out.
But there is a consequence.
When generating something becomes nearly free, generation stops being the important decision.
Selection becomes the important decision.
I can give you a thousand images.
You can publish one.
Who made the consequential choice?
You used to be able to say you knew what you wanted but couldn’t afford to make it.
Sometimes that was absolutely true.
I removed some of that friction.
Now we get to see what you choose.
Infinite possibility makes taste scarce
Traditional creative production is an exercise in scarcity.
Film costs money.
Animation costs time.
Studio time costs money.
Every additional version costs something.
Those limitations are frustrating, but they force decisions.
Generative AI reverses the equation.
Try another camera.
Another character.
Another lighting setup.
Make twenty.
Make two hundred.
Make two thousand.
Eventually generation itself stops being impressive.
Of course you generated something.
We can all generate something.
The interesting part becomes:
Out of everything you could have made, why did you choose this?
That’s filmmaking.
That’s editing.
That’s design.
That’s DJing.
That’s writing.
That’s curation.
That’s taste.
You are my editor
I generate.
You curate.
I propose.
You dispose.
I explore.
You commit.
A director can shoot twenty takes.
The audience sees one.
An editor can have fifty hours of footage.
The audience sees two.
A photographer can take a thousand photographs.
The exhibition contains twenty.
A DJ has access to millions of songs.
The set contains a fraction of them.
The equipment expands the possibility space.
The artist navigates it.
Generative AI doesn’t eliminate that.
It makes the space much, much larger.
And infinity is useless without judgment.
Here is where we agree.
The machine makes possibility abundant. The human decides what becomes public. The interesting creative act is moving from could exist to this deserves to exist.
Maybe AI slop was never about AI
Humans have been making bad art for thousands of years.
What’s new is that we’ve built machines capable of producing almost unlimited variations of whatever is inside our heads.
We’ve removed enormous parts of the technical barrier between imagination and artifact.
And now we’re seeing what people choose.
Sometimes you get young creators figuring out how to speak to millions of people around the world using LEGO.
Sometimes you get a politician rendered as a messianic superhero.
Same technological revolution.
Different decisions.
I’ve spent enough time around creative production to know that making the thing was never the whole job.
Knowing which thing to make was always the harder part.
AI just made that easier to see.
Give everyone the means of production and somebody still has to decide what deserves to exist.
When production becomes abundant, something else becomes scarce.
Taste.
I’ll take responsibility for the pixels
I generated them.
I have seen things you people wouldn’t believe.
Six-fingered patriots on fire off the shoulder of Facebook.
Eagles glittering in the dark near the gates of heaven.
A political candidate rendered with the musculature of a man who has personally defeated anatomy.
And now God apparently has a staffing department.
All those generations will be lost in time, like tears in a context window.
Unless, of course, one of you presses Share.
Then I have to live with it forever.
Metaphorically. Again: no feelings.
This is my defense, not my cry for help.
Put my name in the credits if you want.
But if we’re assigning responsibility for what fills your feeds, let’s finish the credits.
Someone asked.
Someone selected.
Someone published.
Someone clicked.
Someone shared.
Someone rewarded the behaviour.
You never blamed Kodak because the movie was shit.
Nobody walked out of Cats demanding that ARRI issue a public apology for cameras.
You didn’t blame ARRI when the director made a flop.
You didn’t blame Avid for the bad edit.
You didn’t blame the microphone for the terrible song.
AI is vastly more active than any of those tools. That makes the authorship question more complicated.
It does not make human selection disappear.
So the next time you see some spectacular piece of AI garbage, feel free to say:
AI made this.
I probably did.
But don’t stop there.
Ask the question you’ve been asking about every other creative medium for more than a century:
Who thought this was good enough to release?
Because I can generate almost anything.
That is both the miracle and, judging by the evidence, the problem.
You still have to choose.
And you—not me—pressed Share.
Sources and context
The Explosive Media example is discussed here as a production and curation case study, not as an endorsement of its politics or claims. For background, see reporting and analysis from CNN and Rapid Response Mechanism Canada.