Pillar · The dependency lexicon
What is slopaganda?
The short answer
Slopaganda is AI slop weaponised as propaganda: political content mass-produced by machines — fake images, synthetic quotes, bot-fed outrage — that blurs the line between information, persuasion and noise. The word stitches together slop (cheap AI-generated filler) and propaganda (deliberate persuasion). It isn't one deepfake; it's the flood.
This is a reference page, not a daily dose. Read it slowly, then go do something without a screen.
Slop, first
Somewhere over the course of 2024, the internet settled on a word for the rising tide of low-effort, machine-generated content pouring through every feed: slop. Articles no human wrote. Images no one photographed. Reviews no customer left. Comments that exist only because a model was asked to fill a box. The word is deliberately contemptuous — slop is what you pour into a trough.
What defines it isn't that a machine made it; plenty of careful work is machine-assisted. It's that it was made to be made, volume over craft, occupying attention rather than earning it. Slop is content as exhaust.
From slop to slopaganda
Slop is mostly annoying. Slopaganda is slop with an agenda. When the same cheap pipeline that floods your feed with disposable images turns to politics, you get slopaganda: a candidate who never said the quote now attributed to them, a crowd that never gathered for the photograph, a "concerned citizen" comment written ten thousand times by ten thousand accounts that are, in fact, one model humming to itself.
The old propaganda needed a printing press, a studio, a network. Slopaganda needs a prompt and a queue. It is propaganda at the cost of slop — which is to say, almost free, almost infinite, and almost impossible to trace.
Crucially, slopaganda doesn't need to convince you of anything specific. It just needs to fill the channel — to make the true and the fake equally exhausting to tell apart, until the difference stops mattering.
How to spot slopaganda
You won't catch it with a single trick. But a few habits help.
- Provenance is missing, or circular. The claim links to a post that links to a post that links to nowhere. Real reporting names a source; slopaganda dares you to find one.
- The image is too perfect, or too perfectly emotional. The lighting is flawless, the crowd is symmetric, the tear is exactly where you'd want it. Reality is messier.
- Volume without humans. The same take, paraphrased slightly, appears across dozens of fresh accounts in the same hour. People don't coordinate that well; models do.
- The account has no history, or a borrowed one. It joined last week and only posts about one thing, or it used to sell shoes and now exclusively breaks news.
- The claim outruns the source. By the time the correction arrives — if it ever does — the post has been shared a million times. Slopaganda is optimised for the first hour.
- It feels made to be shared, not to be read. The outrage is pre-cut to fit your timeline. If it gives you the perfect thing to say to the people you already disagree with, it was probably made for exactly that.
Why it matters
The danger of slopaganda isn't a single lie. It's the cheap production of plausible noise at a scale that exhausts the very faculty we use to sort signal from noise: our shared sense of what's real. When everything could be fabricated, the tempting move is to believe nothing — which is just as useful to the people filling the channel. Skepticism, weaponised, becomes another kind of slop.
The fix isn't to develop a better eye for fakes (though it helps). It's to slow down, check provenance, and refuse to be a free distribution node for things you haven't verified.
Slopaganda, the hAIroin take
hAIroin is a satirical daily blog about our addiction to artificial intelligence, and slopaganda is one of its recurring fixes. We treat the whole phenomenon as a dependency: we keep reaching for the cheap, the fast, the machine-made, and we keep being surprised when the channel fills with poison. The satire is the needle; this page is the leaflet that comes in the box.
Browse every dose we've filed under the slopaganda tag for the satirical end of the subject. For the wider phenomenon, read what AI slop is. For the worldview behind it, read the manifesto.
Breaking the loop
One practical step, today: pick a single topic you care about and, for one week, refuse to share any "breaking" claim about it unless you can name where it came from. Not the account that reposted it — the original reporting. Most slopaganda dies at that checkpoint.
If you want to go further, the Center for Humane Technology's Take Control guide is a solid, non-commercial starting point for taking your attention back from the machine.
Side effects may include: lucidity, a sudden distrust of perfectly-lit crowds, and the strange urge to verify before you share.