Open Editor
Privacy
6 min read

"AI Slop" Is Merriam-Webster's Word of the Year — Here's What Changed

The word is mocking, not alarmed — Merriam-Webster's editors compared it to the sound of something you don't want to touch. The data behind it is a lot less funny.

In December 2025, Merriam-Webster named "slop" its word of the year, defining it as "digital content of low quality that is produced usually in quantity by means of artificial intelligence." The word itself is centuries old — it meant soft mud in the 1700s, then pig feed and food waste in the 1800s, before settling into a general term for rubbish. Attaching it to AI output wasn't really about fear, according to the dictionary's own editors: it's a dismissive, slightly disgusted label, chosen for having "the wet sound of something you don't want to touch" rather than the tone of something to be afraid of.

Editing a draft and want to make sure it reads clean, not sloppy? ClearText Editor gives you a distraction-free space to write and revise, nothing uploaded anywhere.

The Number That Made the Word Necessary

The timing wasn't arbitrary. According to research from the SEO firm Graphite, AI-generated articles now make up more than half of all English-language content on the web — a genuine milestone, and the kind of number that turns a niche complaint into a word people need. Once AI-written material stopped being the exception and became the statistical majority, "there's a lot of low-quality AI content out there" needed a shorter way to say.

Google Decided to Do Something About It

In March 2026, Google rolled out an update specifically targeting what it calls scaled content abuse — mass-produced, AI-generated pages built to rank rather than to inform. The impact landed hardest on exactly the sites "slop" was coined to describe:

Site TypeTraffic Impact
Niche info sites (500+ AI pages)-60% to -80%
AI affiliate comparison sites-40% to -70%
News aggregators with AI rewrites-50% to -75%
Educational content farms-45% to -65%
Location-based template pages-30% to -60%

Google's detection reportedly didn't rely on flagging AI text itself — it's a multi-signal approach looking at publishing velocity (10 or more articles a day from one domain), near-identical structure across hundreds of pages, missing author credentials, and poor engagement patterns like readers immediately bouncing back to search. Recovery for penalized sites has averaged around six months, and mostly requires consolidating thin pages into fewer, more substantial ones rather than lightly editing what's already there.

There's a distinction the word "slop" doesn't make, and it matters: more than half the English web being AI-generated is not the same claim as more than half of it being slop. A well-edited piece drafted with AI assistance and a mass-produced page nobody reviewed can both technically be "AI-generated," but only one of them is what the word of the year is actually describing. The volume made the label necessary; the quality is what the label is actually judging.

What Separates "AI-Assisted" From "Slop"

The sites hit hardest by Google's update shared a pattern that has nothing to do with whether a model was involved in drafting: no human review at the volume being published, no actual expertise behind the claims, and structure so repetitive across hundreds of pages that it reads as a template rather than writing. None of that is inherent to AI-assisted content — it's what happens when quantity fully replaces editing as the goal.

That's a genuinely different thing from drafting with AI help and then actually reading, fact-checking, and tightening the result before it goes out — which is closer to how most professional writing has worked for a long time, model or no model. The honest-but-unglamorous defense against ending up in the "slop" category is the same thing it's always been: someone has to actually read the draft before it's published, and catch the parts that don't hold up.


The honest summary: "slop" became word of the year not because AI writing exists, but because it crossed a measurable threshold — more than half the English-language web — without a matching increase in editorial oversight. Google's response in March 2026 shows what the correction looks like in practice: traffic drops of 50-80% at sites that mistook publishing volume for a strategy, and months to recover by actually consolidating and improving content rather than just producing more of it.

For questions or inquiries contact us at info@cleartexteditor.com