What is AI slop?
AI slop is low-quality digital content mass-produced by AI — junk articles, weird images, fake news, filler videos. Merriam-Webster named "slop" its 2025 Word of the Year on December 15, 2025, defining it as "digital content of low quality that is produced usually in quantity by means of artificial intelligence."
Why — the first-principles explanation
Slop is an economics story, not a technology story. Publishing content used to cost something — a writer's hours, a photographer's day rate. That cost was a filter. It didn't guarantee quality, but it meant somebody had to believe the piece was worth making. Generative AI dropped the cost of a plausible-looking article to roughly a fraction of a cent and a few seconds.
When production cost collapses toward zero and platforms still pay per view, the optimal strategy flips. You no longer need anything to be good. You need volume. If a thousand articles each earn a tiny amount from search traffic or ad impressions, and each costs almost nothing, the math works even though every single one is worthless to read. Slop isn't a failure of AI — it's AI succeeding at exactly what it was pointed at.
The word choice is doing real work too. Merriam-Webster's president described the flood as absurd videos, weird advertising images, cheesy propaganda, fake news that looks real, and junky AI-written books. Note the tone: mocking, not fearful. Amid years of AI doom talk, the public's most-adopted AI word turned out to be an insult. "Slop" carries its own history — soft mud in the 1700s, pig food in the 1800s, then rubbish generally. Calling AI output slop says it's not scary, it's cheap.
The second-order problem is what economists call a commons collapse. Search results, social feeds, book listings, and review sites all assumed content was expensive enough to be somewhat sincere. Remove that assumption and the signal drowns. And there's a loop: new models train on the internet, the internet is filling with slop, and models trained on model output tend to degrade.
An example that makes it click
Imagine a town where anyone can print a newspaper, and the town pays you a penny for every copy someone picks up. When printing costs a dollar a copy, you write real news — it's the only way to make the penny worth it. Then someone invents a printer that makes copies for a hundredth of a penny.
Now the winning move is to print 10,000 papers full of nonsense with exciting headlines. Each one earns almost nothing. It doesn't matter — you printed 10,000. Meanwhile the one real newspaper is buried under the pile, and the townspeople stop trusting newspapers at all. Nobody set out to destroy the news. They just followed the price. That's AI slop.
Key facts
- Merriam-Webster announced "slop" as its 2025 Word of the Year on December 15, 2025.
- Merriam-Webster defines slop as "digital content of low quality that is produced usually in quantity by means of artificial intelligence."
- The word's older senses run from "soft mud" in the 1700s to "food waste" (pig slop) in the 1800s to "rubbish" generally; the AI sense was coined in the 2020s.
- Merriam-Webster's president Greg Barlow characterized slop as "absurd videos, weird advertising images, cheesy propaganda, fake news that looks real, junky AI-written digital books."
- NIST's Generative AI Profile (NIST-AI-600-1), released July 26, 2024, treats synthetic content proliferation as a distinct generative AI risk category.
- The term is pejorative, functioning much like "spam" — it describes quantity-driven low-value content rather than any specific technology.
▶ The 60-second explainer (script)
AI slop is low-quality digital content mass-produced by AI. Merriam-Webster made slop its Word of the Year on December 15th, 2025, defining it as digital content of low quality produced usually in quantity by means of artificial intelligence. But here's the thing — slop is an economics story, not a technology story. Publishing content used to cost something. A writer's hours. A photographer's day rate. That cost was a filter. It didn't guarantee quality, but it meant somebody had to believe the thing was worth making. Generative AI dropped the cost of a plausible-looking article to a fraction of a cent and about four seconds. So when production costs collapse toward zero, but platforms still pay per view, the winning strategy flips completely. You don't need anything to be good anymore. You need volume. A thousand articles, each earning a trickle of ad money, each costing basically nothing — the math works even if every single one is worthless. Slop isn't AI failing. It's AI succeeding at exactly what someone pointed it at. And notice the word people picked. Not a scary word. A mocking one. After years of AI doom talk, the public's favorite AI term turned out to be an insult. Slop meant soft mud in the 1700s, then pig food, then rubbish. Calling AI output slop says: you're not terrifying, you're cheap. The real damage is downstream. Search results, feeds, book listings, reviews — they all quietly assumed content was expensive enough to be sincere. Take that away and the signal drowns. Worse, new models train on the internet, and the internet is filling with slop.
What authoritative sources say
People also ask
Where did the term AI slop come from?
It grew online in the 2020s as a pejorative, borrowing from slop's older meanings — pig food, rubbish. Merriam-Webster made it the 2025 Word of the Year on December 15, 2025.
Is all AI-generated content slop?
No. Slop specifically means low-quality content churned out in volume. AI writing that someone edited, checked, and actually intended to be useful isn't what the word describes.
Why is there so much AI slop?
Because producing it costs almost nothing while platforms still pay per view. When cost collapses and payment stays, volume beats quality as a strategy.
What is 'workslop'?
AI-generated reports, summaries, and documents that look complete but waste colleagues' time because nobody checked them. Merriam-Webster cited it as one form of the 2025 slop wave.