I spent part of a recent afternoon inside a client's Google Search Console, working through their Pages report. A few hundred URLs sat under "Crawled - currently not indexed." Google had visited each one, looked, and decided none of them earned a spot in the index. A lot of them were fast, cheap, AI-drafted blog posts published to "stay active."
That's the whole story of AI content and rankings in one screen. Google didn't punish those pages for being AI. It just never found a reason to keep them.
So let me answer the question people actually type into the search bar, because the honest answer comes in two parts.
Does Google penalize AI-generated content?
No. Not for being AI. Google has said for years that it focuses on the quality of content rather than how it was produced, and it repeated that plainly when generative tools took off. Its guidance on AI-generated content rewards helpful, original, people-first work "however it is produced." Using AI to write is not against the rules.
Here's the part that trips people up. That same guidance draws a hard line at scale. In Google's own words, "using generative AI tools or other similar tools to generate many pages without adding value for users may violate Google's spam policy on scaled content abuse." Read it closely. The trigger isn't AI. It isn't even volume by itself. It's volume without proportional value. You could publish a hundred thin pages by hand and trip the exact same wire.
Google isn't anti-AI. It's anti-thin. A single AI-assisted page grounded in real expertise is fine. A hundred AI pages built to blanket a topic and catch stray keywords is the thing that gets flagged. Same tool, completely different intent.
So what actually changed in 2026?
AI content has been around long enough that Google stopped treating it as a novelty and started treating it as an enforcement problem. Three moves this year, in order:
Google ran a broad core update from March 27 to April 8, its first of the year. Core updates recalibrate quality broadly and move sites up or down on overall helpfulness, not on any single fix.
On May 15, Google revised its search spam policies to explicitly cover attempts to manipulate its generative AI features, things like buying citations or poisoning recommendations to get quoted in AI answers. That's about gaming AI answers, a different lane from mass-producing pages, but it shows where Google's attention is pointed.
From June 24 to 26, Google rolled out a spam update through its automated systems, including SpamBrain. Google didn't announce a new policy or name a new system. SEO analysts read it as content-focused rather than link-focused, since Google runs link spam as its own separate update. Google itself didn't itemize it, so treat the specifics as informed reading, not an official statement.
None of these is "the AI penalty" people brace for. Taken together, though, they point one direction: Google is spending its enforcement energy on low-value, mass-produced content, and AI made that content cheap enough to flood the zone.
Google is starting to judge patterns, not just pages
For most of SEO history, the game was page by page. You optimized a page, and it passed or failed on its own merits. That's shifting. A 2026 Google Research paper on detecting coordinated synthetic-media abuse describes spotting networks of accounts that share infrastructure and reuse the same generated templates, then acting on the whole cluster at once rather than judging one item at a time.
Be careful how far you take that. The paper is about online video platforms, and it's research-stage, so it is not how your website gets ranked today. But it tells you where the thinking is headed: away from "is this page good" and toward "is this a pattern of low-value production." The volume era, where more pages simply meant more chances to rank, is ending.
I see the small-business version of this in client accounts all the time. A site starts with clean pillars and spokes, everything in its lane. Then the content creeps. Someone writes a post that's a little off-topic, then another, chasing whatever felt urgent that week. The site slowly loses the plot, and Google seems to notice the incoherence. Pages that used to index fine start slipping, and the new tangents never get picked up at all. Drifting out of your swim lane doesn't just waste the new posts. It dulls the good content you already have.
If half the web is AI now, why isn't it all ranking?
Here's the reassuring part, and it's backed by data. By late 2024, Graphite found that just over half of new web articles, about 50%, were primarily AI-generated. You'd expect search results to be drowning in the stuff. They aren't.
In a June 2025 analysis, Graphite found only about 14% of articles ranking in Google were primarily AI-generated. Human-written work took 86%, and an even wider share of the top spots. In AI assistants the gap was starker still: roughly 82% of the content ChatGPT and Perplexity cited was human-written.
These numbers come from AI detectors, which aren't perfect, and they count content that's primarily AI, not any page an AI touched. Studies that count any AI involvement report much higher figures. Different methods, different questions. The direction is what holds up.
The takeaway survives the caveats. A flood of AI content is being produced, and only a thin slice of it earns rankings or citations. Detection is already quietly discounting the mass-produced material before anyone gets around to a manual penalty. Substance is the filter. If you want the fuller picture of how AI engines pick what to surface, that's the whole subject of generative engine optimization.
How to tell if this is happening to your site
You don't have to guess. Open Google Search Console, go to the Indexing section, and open the Pages report. It splits your URLs into indexed and not indexed, and under "Why pages aren't indexed" it lists the reasons with counts. Two labels matter most here.
Google fetched the page and chose not to index it. This is the value signal. When it's stacked with thin or templated posts, Google is telling you plainly that those pages don't earn their place.
Google knows the URL exists but hasn't crawled it yet. On a bloated site, that's often Google deciding your pages aren't worth the crawl, especially when similar pages have been low value.
This isn't rare. While writing this, I pulled the Pages report across ten different Search Console accounts, sites de-identified and ranging from a few dozen URLs to several million. The same pattern showed up in every single one.
If a big share of your published posts is sitting in either bucket, that's your signal. The number-one reason pages don't show up usually isn't a broken tag you can fix in an afternoon. It's Google looking at the page and deciding it doesn't earn a spot. For the deeper version of what these labels mean and how to work through them, I got into why Google deindexes pages separately. And if you're not sure whether AI tools can even see your site to begin with, here's how to tell whether your site is invisible to AI search.
What to do about it, especially with a small team
If you have a real content team and real editorial standards, keep doing what works. This section is for everyone else: the small business, the founder, the one-person marketing department. For you, the volume game is a trap, and here's where I land today.
Cranking out new posts feels productive. It rarely is. Every thin page costs crawl budget, tends to go stale, and often never gets indexed because it doesn't say anything only you could say. You end up spending time or money to make your site harder for Google to love.
The better move is smaller and slower. A manageable set of evergreen pages you actually maintain beats a content treadmill every time. Concretely:
Pull the "Crawled - currently not indexed" list and read it honestly. Those are your problem pages. That's where the work is, not in ten more drafts.
Rewrite it with a genuine point of view, real data, or first-hand experience. Redirect it into a stronger page with a 301 if it overlaps something better. Or retire it. Three options, one decision per URL.
Keep new content inside your pillars and spokes. If an idea doesn't fit a lane, it probably doesn't belong on the site. Topic drift is how good content gets dull.
Pick your handful of important pages and update them with new findings, data, and examples on a schedule. Fresh and maintained beats new and abandoned, every time.
The pushback I hear most: "How am I supposed to answer every question people ask about what we do if I can only publish so much?" Fair question. But that's the quality-versus-quantity choice wearing a disguise. Showing up once for a topic isn't the win. If the page doesn't stay fresh and relevant, that appearance fades, and in AI search it can vanish fast. That's the whole reason an AI citation has a shelf life. One page you keep genuinely current will out-earn ten you abandon. Just because you can publish more doesn't mean you should.
AI content isn't the problem. Publishing without judgment is. Google isn't hunting for AI writing. It's getting better at spotting sites that produce a lot and say little, and it's starting to read that across a whole site rather than one page at a time. If you're a small team, that's good news. You don't need more content. You need fewer pages worth indexing, and the discipline to keep them that way.
Sorting out what to keep, what to merge, and what to let go is exactly what an AI visibility engagement is for, along with how you show up across search and AI answers as one job rather than two. If that's where you're stuck, and you want a second set of eyes on what's actually earning its place, start a conversation. A technical SEO audit is usually where this gets untangled.
Frequently asked questions
No, not for being AI. Google says it focuses on the quality of content, not how it was produced. What it does penalize is scaled content abuse: mass-producing pages that don't add value, whether they're written by AI, people, or both.
Scaled content abuse is Google's spam policy against generating many pages primarily to manipulate search rankings rather than help users. The trigger is volume without proportional value, not the use of AI itself.
Detection isn't perfect, and Google says it doesn't target AI content as a category. But it is increasingly good at spotting the patterns that come with mass production: thin value, templated output, and pages that don't earn engagement or indexing.
AI is fine as a drafting tool. The mistake is using it to publish more. For a small team, a few evergreen pages you maintain and fill with real expertise will outperform a high volume of AI-drafted posts that go stale and never get indexed.
Sources
- Google Search Central, Google Search and AI content (guidance on how AI-produced content is treated)
- Google Search Central, Using generative AI content and the scaled content abuse policy
- Google Search Central, Spam policies for Google web search (updated May 15, 2026)
- Search Engine Land, March 2026 core update and June 2026 spam update
- Google Research, Scalable Detection of Adversarial Synthetic Slop and Coordinated Media Abuse (2026)
- Graphite, More Articles Are Now Created by AI Than Humans and AI Content in Search & LLMs