Search marketers are buzzing about Google’s John Mueller calling AI-generated content “spam.” I did this while answering a question during the April 1 Google Search Central SEO office-hours hangout (and it was no joke).
Mueller’s response is not shocking. Especially considering Google has answered this question many times in recent years. Plus, automatically generated content has long been part of their webmaster guidelines as something to avoid.
Yet, AI-generated content is a popular topic of discussion on social media, forums and in private groups. Especially in recent years as the technology has advanced.
Let’s recap Google’s history on this topic and what it means for you.
The latest Google statement on AI content
First, it’s always important to remember that when Mueller speaks during office hours, he is generally answering specific questions about certain situations. Often, his answers from him get misinterpreted to apply more broadly or mean more than what he actually said. If not blown completely out of proportion.
In this instance, the question was: how does Google react to websites hosting AI-written content?
The answer was straightforward: this falls into the category of automatically-generated content. Which, again, dates back to Google’s earliest days.
But here’s where things get interesting. Can Google tell the difference between content written by AI or a human? Mueller declined to say that definitively.
Ultimately, it doesn’t matter if the content is written by a human or a robot. Google can detect whether it is high-quality, low-quality, or outright spam. At some point, it’s highly likely that Google’s Page 1 results will be filled with content generated by robots.
Heck, for some SERPs it looks like this is already happening. I just did a search for [benefits of AI content] and saw this article ranking in Position 2:
Did a human write that? Or a machine? 🤔
I know plenty of terrible human writers. They could publish content online. But that doesn’t mean Google has to index it or rank it. But anybody can publish any content of any quality online.
Official Google guidance on automatically generated content
Straight from Google Search Central documentation, here is everything Google says about automatically generated content:
Automatically generated (also called “auto-generated”—content) is content that’s been generated programmatically. In cases where it’s intended to manipulate search rankings and not help users, Google may take actions on such content. Some example cases include, but are not limited to:
- Text that makes no sense to the reader but which may contain search keywords.
- Text translated by an automated tool without human review or curation before publishing.
- Text generated through automated processes, such as Markov chains.
- Text generated using automated synonymizing or obfuscation techniques.
- Text generated from scraping Atom/RSS feeds or search results.
- Stitching or combining content from different web pages without adding sufficient value.
More past statements from Google on AI content
2022: Mueller was asked about AI-based content creation tools like Jasper and tweeted back: “Content generators / spinners have been around since the start of the web. People have used all kinds of tools & tricks to do that (see image). As far as I can tell, most sites have trouble creating higher-quality content, they don’t need help creating low-quality content.”
2021: Mueller said that Google would likely shift to focus more on the quality of the content rather than how it was generated. Meaning, essentially, that Google might be OK with ranking machine-written content. But that day hasn’t arrived yet.
2020: In discussing how GPT-3 (Generative Pre-trained Transformer 3, which can produce human-like text and translate content) underperforms, Google’s Gary Illyes said Google doesn’t want machine-translated content in its index. I have shared an amusing example on Twitter: “All boiled together After boiling, the sugar dissolves can be used. Do not let boil for long Coconut milk will be children.”
2019: Mueller was asked whether auto-translating content could lead to a manual action. In short, he said no, but if the content was poor it probably wouldn’t rank well. Also that year, Mueller said machine-written content is not OK. However, I have noted that at some point “down the road” Google might be more open to content generated by a machine. The key would be Google would have to be unable to tell whether it was written by a script or a human.
2017: Illyes was asked whether tools that generate readable content for humans using data are considered “automatically generated content”. Illyes replied that Google was thinking about this but had nothing to say at the time.
2010: Mueller said using automated translation tools (like Google Translate) to create content for your website could, in some cases, be viewed as “creating auto-generated content, which would be against our Webmaster Guidelines.”
An ironic stance?
Google has called itself an AI-first company. They use hundreds of machine learning models in various products – search, Ads, YouTube, Gmail and beyond. why? Because it helps improve the product.
So why is content in search results different? Is it? Well, it all comes down to quality, or at least how Google’s algorithms interpret quality.
But this isn’t an ironic stance really. Google has been consistent about wanting to reward high-quality content. The Panda update was one of Google’s big attempts at cleaning up the mess that content farms had made of search results.
AI content risks vs. benefits
Below are a few risks and benefits of AI content. This is by no means an exhaustive list.
Risk: Google takes action on your page/site
A manual action can have serious consequences for your brand or business. In addition to all the resources you put into this type of content, now you have to sink even more resources into cleaning up the mess and (hopefully) getting back into Google’s search results.
Risk: Not original
While AI content generation is more advanced than the older content spinners, it’s still basically the same thing. You’re copying other people’s work that already exists online – probably including your competitors. So you’re just altering the words and ideas of others. Following, instead of leading, is a dangerous model for any business and pretty much will guarantee you’ll always be number two at best.
Risk: Quality is low
What you get is going to need extensive editing. Best case, you’ll get robotic, vanilla content. With the same (or less) time and resources you invest into technology, you probably could hire a human writer.
Benefit: Writing is time-intensive
It takes humans time to produce content. AI content can cut down this time. That said, you should factor in a resource for proofreading whatever content you get. The Associated Press got some attention for using robot journalists on stats-heavy stories a few years ago, but it increased output while freeing up 20% of journalists’ time.
Benefit: Good content isn’t cheap
And there’s a reason for that. Generally, there’s much more involved with content than the actual writing. There’s research (keywords, what’s ranking, who you’re competing against, etc.). How much editing is required? Are any graphics needed to support it? Is the writer a subject matter expert on the topic? Expertise is at cost. That doesn’t even factor in the promotion of the content. So if you can live with content that is “good enough” (due to limited resources/budget or maybe you operate in a niche with low competition), then AI content could be a benefit for you.
Benefit: Good for idea generation
Writer’s block is the worst. AI can help you come up with ideas for content. At this point, it’s not quite good enough to fully create content from scratch. But maybe there is enough value in the tools to use them for brainstorming content ideas.
why we care
Much of SEO is about weighing risks versus rewards. Google’s stance is if it detects this type of content, you could receive a manual action or be deindexed. That said, AI can help with content creation. But even then, a human layer is needed. Whatever route you go with, always make sure your company or client is comfortable with and aware of any and all risks associated with using AI-generated content.