Tuesday, February 4, 2025
What do Google's Search Quality guidelines say about AI Content?

Google mentions AI twenty times throughout the recently updated Google Search Quality Evaluator Guidelines, last updated on January 23rd 2025 - and it is becoming more and more clear how Google is trying to juggle a pro-AI stance and encourage more people to engage with AI products, and manage the abuse of AI generative content (text and images) as a means to game its systems.
Does Google care if I use AI to write my content?
Technically, Yes.
Page main content
Main content (MC) and supporting content (SC) are terms that have been used by Google in these guidelines for a number of years. the main content of your page is your value proposition, the beneficial purpose, the core topic and reason for that page's existence whether it is text, images, video, or audio.
Google clearly starts that:
The Lowest rating applies if all or almost all of the MC on the page (including text, images, audio, videos, etc) is copied, paraphrased, embedded, auto or AI generated, or reposted from other sources with little to no effort, little to no originality, and little to no added value for visitors to the website.
The key factor in assessing the quality of the MC is how well it helps the page fulfill its purpose and provide a satisfying user experience.
In most cases, quality is reflected in the effort, originality, and skill put into creating the content. For informational pages and those covering Your Money (or) Your Life (YMYL) topics, accuracy and alignment with established expert consensus is crucial.
Rand Fishkin covered this in a Whiteboard Friday in 2016, using the example of granola bars - and how you need to align with consensus in key areas, and show the same (if not higher) levels of knowledge, depth, and originality to warrant inclusion in the index - or to be seen as an alternative result for key traffic driving positions.

So what about content paraphrasing with AI?
Copying, paraphrasing, embedding, or reposting content does not automatically make a page the lowest quality.
A Lowest rating is required when nearly all of the MC (including text, images, audio, or videos) is copied, paraphrased, embedded, or reposted with minimal effort, originality, or added value against the source.
The use of Generative AI tools alone does not determine effort or Page Quality rating.
These tools can be used to create both high-quality and low-quality content.
For example, generating original artwork with significant effort using AI can result in high-quality content, while AI can also be used to produce low-quality content with little effort or originality.
This is why example-led content, with unique experiences that logically can't be copied by others speaks volumes, and will be the future of content moving forward.
AI content disclaimers, and EEAT
A lot of websites have worked to incorporate AI into their content workflows, and believe by being transparent about this process can shield them from any potential repurcussions and maintain user trust.
Its common to include in the authorship section a disclaimer that AI has been used to generate the content, in whole or in part, and then reviewed by person A and person B... But Google an established stance on whether or not these statements enforce or diminish trustworthiness.
On page 46 of the guidelines, Google outlines that content is untrustworthy, and has the lowest quality justification if:
The website terms of use states that "some articles" are generated by artificial intelligence and may have errors or be out of date; there is no indication to which pages this statement applies. The information in this article is not trustworthy and is Lowest E-E-A-T.
And then goes onto explain that if you have a disclaimer to cover responsibility and liability for the use or interpretation of the website content - coupled with the AI disclaimers, and that the user should verify the content's accuracy, Google leads with:
... the articles on this website should be considered untrustworthy and Lowest E-E-A-T
Scaled content abuse
Scaled content abuse refers to the practice of generating large amounts of low-quality or unoriginal content, primarily to manipulate search rankings.
Creating an abundance of content with little effort or originality with no editing or manual curation is often the defining attribute of spammy websites.
These can include:
- Doorway pages.
- Duplicate or near-duplicate pages.
- Programmatic content (if executed poorly).
- Mass AI-generated content.
This section does cover the areas listed above, but specific to AI use for scaled content:
Using automated tools (generative AI or otherwise) as a low-effort way to produce many pages that add little-to-no value for website visitors as compared to other pages on the web on the same topic.
Regardless of how the content has been scaled, Google's guidelines for a long time have been clear that "pages and websites made up of content created at scale with no original content or added value for users, should be rated Lowest".
This includes AI content, and even if it is strongly suspected that generative AI tooling has been used to generate mass (poor) content at scale, the suspicion is enough to warrant a score of "lowest".
Even if you are unsure of the method of creation, e.g. whether or not the page is created using generative AI tools, you should still use the Lowest rating when you strongly suspect scaled content abuse after looking at several pages on the website.
When it comes to YMYL topics, Google's stance is strict:
This is spam (scaled content abuse) and a highly untrustworthy medical page: it demonstrates no real expertise and may not be correct and is especially concerning for YMYL topics.
The medical page is a reference to the content example, which is a poor page talking about cancers.
Page FAQs
A common method for generating scaled content involves gathering questions from sources like Google Search's "People also ask" feature and answering them with unoriginal, paraphrased content, often produced using generative AI tools.
These again, are rated as "the lowest" by Google.