What Is "AI Slop" — and Why the Internet Is Full of It

You have definitely seen AI slop. The surreal images of Shrimp Jesus on Facebook. The SEO articles that answer every question by restating it. The AI-narrated YouTube videos with 14 ads and nothing to say. The Amazon listings written by bots, for products made by bots, reviewed by bots. Slop is not a joke — it is now a measurable fraction of the content you encounter every day online.

What does "AI slop" actually mean?

The term was popularised by writer Simon Willison in 2024, and quickly went mainstream. Slop refers specifically to low-effort AI content — created in volume, optimised for algorithms rather than humans, with no care for quality or accuracy. Not all AI content is slop. A carefully edited AI-assisted article is not slop. The defining feature of slop is laziness: generated in seconds, published without review, made only to capture attention or ad revenue.

Where do you see AI slop most?

• Facebook — bizarre AI images ("baby boot camp", "log cabin made of lemons", "Shrimp Jesus") that rack up millions of engagements from older users who do not realise they are fake
• Google Search — AI-generated SEO articles that answer questions by restating them, dressed up with fake author photos
• YouTube and TikTok — AI-voiced videos on "top 10" or "facts about" topics that recycle the same Wikipedia text over and over
• Amazon — product listings written by AI, photos generated by AI, reviews written by AI. Sometimes the product itself does not really exist
• LinkedIn — generic "motivational" posts and "thought leadership" articles that are obviously AI-drafted, clogging feeds
• News sites — entire outlets where every article is AI-written with no editorial oversight

Why is there so much of it?

Because it is cheap, fast, and — until recently — paid. Writing a blog post used to cost $50 to $200. With ChatGPT, it costs less than a cent. If even a tiny fraction of those AI-written pages get ranked on Google, clicked on, or shown ads, it is profitable. The same applies to AI images on Facebook, AI videos on YouTube, AI listings on Amazon. Platforms monetise engagement, and slop generates engagement at almost zero cost.

How do you spot AI slop?

• Images — impossibly symmetrical, too-smooth skin, warped hands, text that does not parse, lighting that is too perfect
• Articles — every paragraph opens with a hedge ("It is important to note"), no specific examples, vague claims, no author information
• Videos — robotic or slightly-off voiceovers, generic stock footage, no personality, descriptions written like product spec sheets
• Product listings — odd phrasings in titles, generated-looking photos, reviews that all sound the same
• Social posts — too much emoji, overly formal tone, "hot takes" with zero actual opinion

What the platforms are doing about it

Slowly, and imperfectly. Google now penalises thin AI content in search rankings. Meta started labelling AI-generated images on Facebook and Instagram. YouTube requires disclosure for AI-altered content. The EU AI Act requires labels on AI-generated content across most platforms. But the incentives cut the other way — slop is profitable, platforms make money from it, and enforcement is inconsistent. Expect the slop flood to keep rising through 2026, and for the "dead internet" feeling to get stronger before it gets better.

Bottom line

AI slop is the defining internet phenomenon of 2026. It is cheap, massive, profitable, and — for most platforms — nobody's top priority to fix. Your best defences are awareness and scepticism: notice when something feels generated, check for author bylines, prefer trusted sources, and assume that strangely viral content on Facebook is probably fake. The open web is getting noisier, but good content still exists — you just have to look for it more carefully.