Search bar on computer screen

Photo by Markus Winkler on Unsplash

You know the feeling. You search for something on Google, click a result, and immediately your brain registers that something is off. The writing is technically correct but somehow empty. The article answers your question but says nothing. The information is present but the insight is absent. You’ve just encountered AI slop, and it’s everywhere now.

According to Originality.ai’s ongoing study, AI-generated content peaked at 19.56% of Google search results in July 2025. That’s nearly one in five top-ranking pages created not by humans with expertise but by language models trained to sound like humans with expertise. The current figure has dropped slightly to 17.31%, but the trend established a new baseline for what internet content looks like.

The SEO industrial complex figured out the obvious: AI can generate content at scale for pennies. Why pay a writer $200 for an article when ChatGPT produces something “good enough” for fractions of a cent? The math was irresistible. The flood was inevitable.

What’s happening to search quality is a tragedy of the commons playing out in real-time. Individual publishers gain short-term advantage by producing AI content that ranks. Collectively, they’re degrading the information ecosystem that made search valuable in the first place.

Google’s response has been reactive rather than proactive. They’ve updated their helpful content systems to better identify and demote AI-generated material that adds no value. But detection remains imperfect, and the cat-and-mouse game continues. For every algorithm update that suppresses AI slop, content farms adjust their prompts and workflows to evade detection.

The categories most affected are predictable: product reviews, how-to guides, health information, financial advice. These are high-volume search categories where generic answers satisfy algorithmic requirements even when they fail human needs. When you search for “best running shoes 2025,” you’re increasingly likely to find AI-generated listicles that have never touched a running shoe.

The irony is that Google itself deployed AI in search through SGE (Search Generative Experience), now simply called AI Overviews. Users increasingly see AI-generated summaries at the top of results, followed by organic results contaminated with AI-generated content. The human-written, expertise-driven content that made the web useful is being squeezed from both directions.

What can users do? Append “reddit” to searches to find human discussions. Look for bylines and author credentials. Check publication dates and update frequencies. Develop familiarity with sources you trust. These are workarounds, not solutions, but they’re the reality of navigating 2025’s degraded search landscape.

The 17.31% figure likely understates the problem. Originality.ai’s detection has limitations. Sophisticated AI content that’s been edited by humans or generated through careful prompting may evade classification. The true proportion of AI-influenced content is probably higher.

We’re watching the web transform in real-time. The abundance of text that made the internet feel infinite is becoming an abundance of generated text that feels hollow. The challenge for the next phase of search isn’t finding information—it’s finding information that was actually created by someone who knew something.