Google is cracking down on AI-generated junk pages

Google has tightened its grip on websites that use artificial intelligence to generate large volumes of low-quality pages designed mainly to rank in search results. Starting in March 2024, the company strengthened its policy against scaled content abuse and has continued to enforce it. The company later said it saw 45 percent less low-quality, unoriginal content in results. If you've been using AI to mass-produce pages for your website, this matters to you.

Key facts

• Google strengthened its scaled content abuse policy on March 5, 2024, to cover pages created by automation, humans, or both.
• In September 2026, John Mueller from Google stated that programmatic SEO of certain kinds often results in spam or low-quality sites, and that existing pages may continue to influence how Google evaluates the site as a whole.
• Google said it saw 45 percent less low-quality, unoriginal content appearing in search results following these enforcement efforts.
• Websites using AI to generate thousands of pages without adding real value are now facing manual penalties or algorithmic suppression.

What counts as the kind of AI publishing Google penalizes

The line Google draws is simple: Does each page add something real that a visitor actually needs? If you're using AI to swap out a city name or product category and call it a new page, but the underlying information is identical or made up, that's what Google targets. Pages built from guesses or generic templates with no actual data behind them fail this test. Google does not care whether a human or a machine wrote the words. It cares whether the page solves a real problem for a real visitor.

What kind of AI-generated pages still work

Legitimate programmatic publishing still exists and still ranks. The difference is that each page is built on actual data. Think of a currency converter that generates a page for every currency pair, pulling live exchange rates. Each page is unique because the data is unique. Or a cost-of-living guide that pulls rent, transport, and grocery prices from a maintained dataset and updates when numbers change. These pages deserve to exist because they answer a specific query with real information. If you can answer the question "Where did this number come from and when was it last updated," you're following the model that has survived Google updates.

What you should audit on your own site right now

Go through your pages and ask yourself: Would a visitor ever land on this page on purpose if they came to my site directly? If the answer is no, it's at risk. Check whether the information on each page comes from a real source or a real dataset. Look for pages that are nearly identical except for one swapped variable. See whether pages are actually being visited and whether visitors are doing anything useful when they arrive. If you've published hundreds or thousands of pages in the last year or two using AI, start by reviewing the ones that seem weakest or most similar to each other.

What to do if you have pages like this

Delete the ones that don't pass the test. This is not a gray area where you can wait and see. Pages that exist only to capture search traffic and offer no real value can affect how Google views your site. If you have a system that keeps publishing pages automatically, add a human review step before anything goes live. Have someone who knows your business check whether the information is accurate and whether the page actually answers a search query that real people make. Slow down your publishing. The teams that have survived Google updates are the ones that built a rule for deleting pages before they needed one.

Sources

• Why the data center you dread might never even be built - TechCrunch, November 11, 2026
• Good Vs. Bad pSEO: Learnings From A Decade Of Building Scalable Content Engines via @sejournal, @_nitman - Search Engine Journal, November 10, 2026

They did the reporting. The writing and the opinions here are ours.