Are you a fan of AI-generated content?
Then you must have wondered whether Google penalizes AI content or not.
Google has now clarified that they are not against AI content. And in short, this means there is no penalty for AI content.
But still, I experience many being wary of whether Google will detect their AI content and penalize it. Even if Google say they won't.
And so many SEO professionals are still hesitant to use AI, fearing that it will be penalized by Google.
So let's look at the hard facts, cases and the reason why so many believe Google will downrank their content.
What is AI-generated content?
AI-generated content is content created by artificial intelligence software rather than human writers.
This technology has become increasingly popular in recent years, allowing businesses to create large amounts of content more quickly and easier than before.
But a misconception I often meet is that AI content equals walls of text or plain text articles.
This is not the case.
AI content can cover a variety of different elements suchs as:
- AI-text: articles, stories, poems, product descriptions, chatbot responses, etc.
- AI-generated media types: photographs, illustrations, infographics, music, speech, sound effects, animated videos, commercials, film scenes, etc.
- AI-voice: speech for virtual assistants, narration, voiceovers, etc.
- AI-generated Code and Programming: software, apps, games, etc.
- AI-generated Financial Data and Predictions: stock predictions, financial reports, risk analysis, etc.
- AI-generated Scientific Data and Predictions: weather forecasting, scientific research, drug discovery, etc.
In another of my articles, I outlined how I used a combination of the following AI methods to make the article. Does this A) make my article AI content, and B) would it be fair that Google should penalize it?
- generated a headline that I then tweaked
- created images for illustration
- benchmarked against competitor
- summarised other old articles
- found relevant topics to include
- wrote part of the text
- did spell-checking
- Rewrote bad written sentences
- identified outbound links
Yes, fair enough Daniel you might think. This is enhanced AI content of high quality. But what about the content that is mainly or only generated by AI and then published? Google must, for sure, downrank that.
Well, I believe this is not the case per see. It depends on the quality of the content, just like Google states.
"Our focus on the quality of content, rather than how content is produced, is a useful guide that has helped us deliver reliable, high-quality results to users for years."
Source: Google Search Central
Let's explore some cases to verify this.
Case examples with AI Content and their performance
When evaluating AI content cases and whether it gets penalties from Google, we must choose what we want to determine as AI content.
In this article, we will look at AI content examples that is either:
- easily identifiable text by some of the AI detectors that analyse whether a text is human or AI written.
- text and content that is known to be AI content.
What does a penalty from Google look like?
Well, it can be a decrease in rankings or even complete removal from search results.
Both can be difficult to identify. So for our case reviews, we will instead reverse it and see if we can identify content that actually ranks and generates traffic.
Case 1: Pages ranking on Technical SEO term
If we look at the search phrase "technical SEO" in Danish (here it's "teknisk seo"), we know for a fact that 2 out of the 10 are mainly generated with AI.
One of them, we are told, is created with AI (ChatGPT) and the other is made with the SEO.ai platform.
And the text is easily identifiable as AI created content by OpenAIs AI text classifier/detector.
And with it being a classic SEO term and something SEO agencies want to rank on, the competition on these terms is also decent.
So first case example of AI content performing without clear penalty.
But Google might not be able to identify or not have gotten to the Danish language yet you might argue. And both are valid points.
In general, AI detectors have a harder time identifying other languages than English (they have less training data in these languages), and Google's algorithm and spam preventions are sometimes less strict/rolled out in non-English languages.
Let's move on to more English cases then.
Case 2: AI-generated content on "Automatic Writing tools"
A more classic example of AI content is our own page targeted the keyword "Automatic Writing tools".
In this case, it ranks #10 for the keyword terms. Not super great, but it a page 1, nevertheless.
And even though the SEO score of the page might be good (although the title is not optimised with best practices), my guess is the lower ranking has more to do with it not meeting the search intent entirely and the user might feel they get met with a wall-of-text instead of a simpler digestible listicle with different tools.
And there are almost no internal links for the page, so it could be something we could improve as well.
So the low number - albeit being on page 1 - is not something I would deem a penalty from Google. It's just their algorithm working and not evaluating the page as the best result for the searchers. I might tend to agree.
Case 3: Bankrate - an already established reputable website
One of the more prominent and well-known examples of AI content at scale is what Bankrate.com published.
The site uses AI content to answer questions with explicit constraints, such as "What Is Contribution Margin" or "What Is Arbitrace"
According to Bankrate.com, subject matter experts review and edit the content before publishing.
If we try to identify what pages are made with assistance of artificial intelligence, a Google search results in 309 pages.
If we take a couple of these pages and review them in Ahrefs, it's clear that they rank for many keywords. The three pages I checked ranked on between 74 and 188 keywords. And even though many of the rankings are low positions (there was also a couple of #1's) or the keywords have low volume, they display a long-tail keyword strategy.
I have seen estimates that the site drives hundreds of thousands of visits from its AI content every month, indicating that AI content performs well for SEO when done right.
The success of AI-generated content for SEO on sites like Bankrate might be attributed to a combination of factors.
One of these factors is the authority of the domain. The site has been around for decades, accumulating strong backlink profiles and establishing itself as a reputable source of information.
This means that Google trusts its content and is more likely to rank it higher in search results.
Additionally, these sites utilize subject matter experts to review and edit their AI-generated content before publishing. This ensures that the content is fact-checked, corrected, streamlined, and sometimes slightly rewritten before it goes live.
Bankrate also follow the Google policy on telling the users who made the content - and let them identify it as made by an AI (if they click the link and look for it. I doutb many will do this)
Also the content created by AI is targeted questions with clear constraints, making it easier for generative AI to create informative and accurate answers. This is known as functional content. The boundaries of the answer are clear and easy to create, which makes it more likely to rank well for relevant keywords.
Case 4: 30,000,000 Organic Clicks in 4 months with AI content
But what about AI content on newly created domains? Can that perform?
I have not yet seen many public cases with AI content on new domains. Or at least now where the domain is disclosed.
But this example from Koray Gübür seems to demonstrate that it is indeed possible.
One might argue that a growth like this will make the Google bot pay more interest to the site to identify whether or not it's spam in general.
Whether this, in the long run, will lead to a penalty for a site like this is clear to determine without knowing the site. Will reach out to Koray to ask what the status is today.
More cases
Now I covered 4 cases with AI content, but there are many other public examples of AI content, including Cnet and Creditcards.com.
And then there are the older ones that are more programmatic SEO (and not built on the newer large language models AIs like OpenAI are providing). Examples of these are The Washington Post and its AI tool called Heliograf or Forbes and it's Bertie. Both facilitate the creation of news articles.
Does Google penalize a single piece of content/page or whole sites?
Neither a single piece of AI-generated content nor a whole site is automatically penalized by Google.
Like any other content, the search engine's algorithm evaluates its quality and relevance to determine its ranking on SERPs.
However, given the potential for abuse and spam by using AI to crank out low-quality content, Google might be more vigilant in scrutinizing AI content.
Read my full article on Google's guidelines on AI content and their most recent statement.
Will google penalize AI content in the future?
As an AI company themselves, I find it highly unlikely that Google will suddenly start penalizing AI content just because it's made by artificial intelligence.
In fact, AI-generated content is becoming increasingly prevalent across the web and is likely to become even more embedded in everything we do.
Google's algorithm is constantly evolving to understand better and evaluate content quality and relevance, regardless of whether AI or humans generated it.
But Google will need to step up its game to evaluate what content to rank, as the amount of content will most likely go through the roof.
And while there is always a risk of spam and abuse with any content generation, Google will continue to prioritize high-quality, relevant content in its search results, regardless of how it was created.
That's why several other factors for ranking will gain prominence and receive increased attention, such as site authority, user signals, backlinks, etc. But I will leave this for another article.
What are Google's Guidelines on AI content?
We covered Google's Guidelines and statements around AI texts in various articles. Especially Google's recent blog post on Google's Developers Blog.
And even though there are still many sceptics or people now being up to date with the latest announcements from Google and still referring to old outdated John Mueller quotes, the search behemoths policies are quite clear.
In short, it's like this;
Google has clearly stated that they reward high-quality content regardless of how it is produced.
While they do not penalize a single piece of AI-generated content or a whole site automatically, they scrutinize AI content for potential spam and abuse.
According to Google's guidelines, AI-generated content should:
- focus on meeting search intent
- adding a human touch
- enriching the text
- avoiding techniques that result in low-quality content
.svg)
