THE PEOPLE WHO NEVER EXISTED

AI Propaganda Meets the AI Backlash

Something strange happens when a propaganda machine invents ordinary people at exactly the moment ordinary people are getting sick of artificiality.

In Pakistan, a short video circulated ahead of a planned PTI march. Young men and women appeared to warn people against violence, vandalism and blocked roads.

Reasonable message.

Except the people weren't real.

Geo Fact Check concluded the video was AI-generated, pointing to recurring faces and visual defects. What remains unknown is who created or commissioned it.

That uncertainty matters. But the technique matters more.

This wasn't a fake president or celebrity. It was something subtler:

manufactured public opinion.

Instead of saying, “The government thinks this,” synthetic media can imply, “people like you think this.”

That changes propaganda.

You no longer need actors, locations or even willing participants. You can generate a believable group of ordinary citizens who appear to agree with whatever message you want distributed.

At the same time, something else is happening.

People are becoming increasingly suspicious of synthetic culture.

“AI slop” has become an insult. Handmade art, bad photocopies, physical zines, film photography, stop-motion, imperfect typography and visible human labor are gaining new cultural value.

The irony is hard to miss.

The more machines produce perfect images, the more imperfection begins to look authentic.

A crooked drawing says somebody made this.

A bad photocopy says somebody touched this.

A strange local zine says somebody cared enough to print fifty copies.

Authenticity itself is becoming scarce.

And scarce things become valuable.

That doesn't mean human-looking media will remain easy to identify.

AI will get better at copying the very flaws we now associate with real people: shaky cameras, poor lighting, hesitation, awkward pauses, imperfect faces, cheap microphones.

Eventually, AI won't just imitate humans.

It will imitate the signals we use to prove something is human.

That's where this gets interesting.

I've also been making plenty of AI art myself, and some of the propaganda aesthetics are ridiculously fun. The giant typography, symbolic exaggeration and theatrical certainty are great creative tools.

So I'm not interested in pretending AI is artistically worthless because it can also be manipulative.

Both things are true.

AI can be an incredible creative medium.

AI can also make it much harder to know who actually said something, who actually appeared in an image, or whether the apparent crowd ever existed at all.

Maybe the backlash is simply culture correcting itself.

Commercial radio produced pirate radio.

Corporate publishing produced zines.

Digital perfection may produce a new demand for things that feel unmistakably human.

And that may be AI's funniest unintended consequence.

Machines become better at perfection.

Humans start making things rougher, stranger and more personal.

Meanwhile, somewhere on the internet, a crowd of people who never existed is telling us what everybody supposedly thinks.

FMPU

The first wave of AI propaganda tried to make machines look human.

The next wave may try to make machines look imperfectly human.

And that is where propaganda and the AI backlash finally collide.