The History of Google’s Relationship with AI
Google and automation have always lived in harmony.
“We’ve long believed in the power of AI to transform the ability to deliver helpful information.” – Google Search’s Guidance About AI-Generated Content, Google Search Central Blog
Its relationship, like all relationships, evolves over time, however. In this post, we unpack the evolution of Google’s relationship with automation and artificial intelligence (AI), as well as detail our own experience and insights about how to best approach automation when it comes to content creation and SEO.
The Fundamentals
Despite keeping many of the nitty gritty details of its algorithm under wraps, one thing Google repeats time and time again is the message that valuable, original content is king.
Create Insightful, People-First Content
In its SEO Fundamentals documentation (an excellent guide for anyone just getting started with SEO), Google states that its ranking system is “designed to present helpful, reliable information that’s primarily created to benefit people.” It then goes on to detail best practices in creating said helpful, reliable, people-first content.
Highlights include:
- Present original information, reporting, research or analyses. This may include gleaning insights, opinions, or salient themes from other sources, not simply restating the findings of others’ work.
- Make sure the content is complete and substantial. Don’t leave the reader hanging with missing pieces or leave them to make assumptions or connect the dots on their own. Give them every detailed piece of information on your particular, specific topic in a single source.
- Provide insightful analysis or interesting, thought-provoking information. Avoid repeating the same information or conclusions that a large number of other content creators have already communicated on a mass scale. Identify some original thought or analysis to add to the mix to push the thinking on a particular subject forward. Ideally, any piece of content should demonstrate aspects of experience, expertise, authoritativeness, and trustworthiness (what Google refers to as E-E-A-T).
Understand the Basic Rules Around the “How” and the “Why” of AI for Content Generation
This same guide references best practices when it comes to the use of artificial intelligence (AI) or automation to produce valuable content.
How to Use AI Appropriately for Content Creation
Google doesn’t specifically direct content creators not to use AI tools, but it does urge them to avoid extensive automation to produce content on multiple topics.
The guide also advises users to provide disclosures about how AI or automation was used to produce the content. For example, was automation used to generate background information on the topic or even to generate the first draft of a piece of content? If so, make that information known to readers so they understand how the content was created.
A Major Mistake to Avoid When it Comes to AI-Generated Content
The one hard and fast rule Google does specifically outline as a big “no no” in its guide is using AI for the sole purpose of ranking in search engines. Google explicitly states:
“If you use automation, including AI-generation, to produce content for the primary purpose of manipulating search rankings, that’s a violation of our spam policies.” – Creating helpful, reliable, people-first content, Google Search Central Blog
Google’s “SpamBrain” system is used to flag instances of ranking manipulation to reduce the amount of low-quality content in search results. SpamBrain uses machine learning and manual enforcement to identify sites with auto-generated and scraped or duplicate content reduce the amount of low-quality content on the web.
The Specifics
Back in February 2023, Google released in-depth guidance about how to use AI-generated content. In this guide, they repeat many of the same tenets in their foundational documentation described above. They reiterate that they reward original, high-quality content that demonstrates E-E-A-T principles. They also repeat ways in which automation can be helpful in creating valuable content but warn creators about using automation for the sole purpose of manipulating search rankings.
Not All AI-Generated Content is Spam
The use of AI on its own to produce content is not against Google’s guidelines. Google says is takes a “responsible approach” to deciphering where automatically generated content lands in its search engine, continuing to reward quality information and helpful content — no matter how it was produced.
AI has the ability to power new levels of expression and creativity, and to serve as a critical tool to help people create great content for the web. – Google Search’s Guidance About AI-Generated Content, Google Search Central Blog
Using AI to produce content neither gives the content a leg up in search, nor does it necessarily get penalized. Where your content is ranked is based purely on the value of the content produced, not by how it was produced.
How to Display Author Bylines When Using AI
Going back to the “How” that Google outlines in its foundational documents, the author byline can be tricky when AI tools are used. Google advises listing accurate author bylines. So, when a hybrid human + AI approach is used, Google recommends including a disclaimer that AI was used to generate the content. On the other hand, Google advises against giving the name of the AI tool or simply “AI wrote this content” in the byline to avoid obscurity when it comes to authorship and the content creation process.
The Evolution
On March 5, 2024, Google announced a core update and new spam policies designed to improve the quality of search. The new spam policies build on its policies about automatically-generated content.
Core Update
Here are key takeaways from this latest update related to the core ranking system:
- For the first time, Google’s core update was tied to its “helpful content” update. Previously, these updates happened independently, with the first instance of the helpful content system rolling out in August 2022. Moving forward, it appears that core updates and helpful content updates will be more closely intertwined, which could lead to greater fluctuations in search rankings since more conditions are impacted simultaneously.
- Core ranking systems were enhanced to include “a variety of innovative signals and approaches” on top of its already robust list of existing ranking signals. What exactly those new signals and approaches are, Google keeps fairly tight to the chest (or I just couldn’t find them in the wealth of documentation). However, they, once again, point to best practices for creating helpful, reliable, people-first content. (So, nothing new there).
- One detail noted with this update is that the core ranking systems are designed to work on the page level, although some site-wide signals are considered. What that means is that a helpful, valuable page on an otherwise “unhelpful” site can rank.
New Spam Policy on Scaled Content Abuse
One of the three new spam policies outlined in Google’s latest update is related to scaled content abuse – or the practice of generating large amounts of unoriginal, unhelpful content for the primary purpose of manipulating search rankings.
This new policy builds on our previous spam policy about automatically-generated content, ensuring that we can take action on scaled content abuse as needed, no matter whether content is produced through automation, human efforts, or some combination of human and automated processes. – What web creators should know about our March 2024 core update and new spam policies, Google Search Central Blog
Again, there’s nothing new here about Google’s stance on AI. There’s no harm in using it, as long as it produced helpful content at the end of the day. It does appear, however, that producing large volumes of content, which are most likely generated using automation of some kind, may put a target on your back.
Here are specific examples of scaled content abuse:
- Generating high volumes of low- to no-value pages at a time, either using generative AI tools or by scraping and synonymizing content.
- Stitching together content from different web pages.
- Creating keyword-rich pages that are non-sensical and unreadable to a site visitor.
Our Findings So Far
What’s the likelihood of getting a manual action from Google if I’m using AI to generate content?
Very low. Google doesn’t care how you generate content. At worst, your content just won’t get indexed if Google deems is unhelpful, but you will not likely have a manual action against your site just because you’re publishing low-quality content (AI-generated or otherwise).
From what we’ve seen, anecdotally at least, is that only the worst cases of spammy sites, such as those with fake buttons, overwhelming ads, suspicious redirects, or websites that are clearly deceptive and offer no real value, are flagged and have manual actions taken against them for violating Google’s spam policies, as they should.
Can AI-generated content rank in search engines?
Yes…For now, at least.
Case in point: as part of our SEO strategy and setup, we create editorial calendars complete with relevant, high-opportunity keywords for our clients. We’ll admit that some of our clients have taken that editorial calendar and simply used ChatGPT or Jasper to generate content using that list of keywords. If that’s not “using AI with the intention of manipulating search rankings”, I don’t know what is. Yet, none of these clients were punished by Google for doing so. In fact, much of their AI-generated content is indexing just fine.
We certainly do not advise to go directly against what Google has explicitly stated NOT to do, of course. As with anything, use this approach with caution and care. Here’s our thinking on this topic:
It’s okay with start with a target keyword in mind first and write content with that keyword in mind. It’s also okay to use AI to generate at least the first draft of your content to give you a head start. What we do advise, however, is to: 1. approach the topic surrounding that target keyword with some original angle or idea, and 2. review, finesse and add value and originality to AI-generated drafts.
It’s largely unknown how well Google can really identify the true intent behind a piece of content, but if the output is of value, you should have nothing to worry about.
Tips on How to Use AI to Generate Content
After taking a deep dive into Google’s foundational philosophy, detailed search algorithm updates, and our own client data, our key takeaways on the use of AI to generate optimized content are this:
- It’s okay to use AI to generate content. Google does not care how you create content. What matters is the quality, originality, and helpfulness of that content.
- Avoid using AI to generate content in large volumes. Google’s new spam policies describe scaled content abuse as the practice of “creating large amounts of unoriginal content that provides little or no value to users.” What “large amounts” means exactly is a bit nebulous, but if you are publishing content so quickly that it’s clear you aren’t reviewing it first, that’s probably too much. The same might go for the production of say 100 landing pages that all say the same thing, for the most part, but target different cities where your business isn’t located.
- Steer clear of low-quality, low-value content. Don’t just add to the noise. When in doubt, use Google’s own guide on how to create helpful, reliable, people-first content.
- Avoid duplicate, unoriginal content. If you are using AI tools to produce content, be sure the copy is original, not redundant with others’ work already indexed in search, to give yourself the best fighting chance of getting your web page or content ranked.
At the end of the day, whether you’re using AI-generated content, real human brain power, or a combination of both, Google’s stance is not really all that far off from its foundational belief: quality content wins.
