Google is directing its quality raters to watch out for pages with main content created using automated or generative AI tools – and rate them as lowest quality, according to Google’s Senior Search Analyst and Search Relations team lead John Mueller, speaking at Search Central Live in Madrid.

This was shared by Aleyda Solis today on LinkedIn:

This change was part of the January 2025 update of the Search Quality Rater Guidelines. In case you missed anything else from that update, here’s a recap of the most significant changes from the latest version.

1. Google introduces a new definition: Generative AI

With its latest Search Quality Rater Guidelines update, Google added a definition and framing for generative AI for the first time. Google’s document calls it a useful tool, but one that can be abused.

The addition of Generative AI appears in Section 2.1 (Important Definitions):

“Generative AI is a type of machine learning (ML) model that can take what it has learned from the examples it has been provided to create new content, such as text, images, music, and code. Different tools leverage these models to create generative AI content. Generative AI can be a helpful tool for content creation, but like any tool, it can also be misused.”

2. Google reorganizes and expands spam definitions

Google significantly overhauled how spammy webpages are defined.

The previous section 4.6.3 (Auto-generated MC) is gone. In its place, Google added new subsections and increased its focus on scaled, low-effort content, including potential AI misuse.

What’s new in 2025 brings the guidelines in line with Google’s big search quality changes from last year:

• Expired Domain Abuse (Section 4.6.3): This happens when “an expired domain name is purchased and repurposed primarily to benefit the new website owner by hosting content that provides little to no value to users.”
• Site Reputation Abuse (Section 4.6.4): When “third-party content is published on a host site mainly because of that host’s already-established ranking signals, which it has earned primarily from its first-party content. The goal of this tactic is for the content to rank better than it could otherwise on its own.”
• Scaled Content Abuse (Section 4.6.5): Creating a lot of content “with little effort or originality with no editing or manual curation.” Generative AI is mentioned as one example of an automated tool used for this.
• MC [Main Content] Created with Little to No Effort, Little to No Originality, and Little to No Added Value for Website Visitors (Section 4.6.6): This is a new catch-all section for low-quality paraphrased content, often seen with generative AI and other forms of automated generation.

Section 4.6.6. is what Mueller called attention to in his presentation, specifically this part:

“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. Such pages should be rated Lowest, even if the page assigns credit for the content to another source.” [emphasis added].

Now, how exactly would a rater know whether content is auto or AI-generated? There is no guidance specific to AI-generated content, but there is some new guidance around “paraphrased content”:

• Section 4.6.6: “Automated tools can also be used to create paraphrased content by restating or summarizing the content on other pages.”
• Section 4.6.7: “Paraphrased content can be much harder to recognize… Paraphrased content is likely to:
  • Only contain commonly known information or generally known facts
  • Have high overlap with webpages on well established sources such as Wikipedia, reference websites, etc.
  • Appear to summarize a specific page such as a forum discussion or news article without any added value
  • Have words or other indications of summarizing or paraphrasing generative AI tools, such as words like ‘As an AI language model’”

3. Google explains low vs. lowest rating

This new section introduces rater guidance for when content isn’t bad enough to get a Lowest rating, but still deserves a Low rating. Here’s the difference:

• Low: Some MC is reused, but there is at least minimal effort to curate or modify it.
• Lowest: Almost all MC is copied or paraphrased with no effort or added value.

The Search Quality Rater Guidelines share examples of repackaged content like:

• “Social media reposts with little additional comment or discussion”
• “Pages with content from other sources (e.g. pages of embedded videos or pages with ‘repinned’ images) with little additional comment, discussion, or curation by the content creator of the page”
• “‘Best’ lists based on existing reviews and lists with little original content.”

Google wants raters to flag thin content that tries to pass as original but doesn’t meet the standard for a quality user experience​.

4. Google adds ‘filler’ content

This new section addresses “filler” content — that is, low-effort, low-relevance content that may visually dominate a page while failing to support its purpose.

“Filler can artificially inflate content, creating a page that appears rich but lacks content website visitors find valuable.”

It emphasizes that even if content isn’t harmful, it can earn a Low rating if it makes it harder to access truly helpful material. Especially targeted: pages that bury useful info beneath ads, generic introductions, or bloated paragraphs:

• Filler that’s prominently placed and distracts from the MC
• Pages that appear longer or richer than they are by padding out space

Raters are encouraged to evaluate how page layout and content hierarchy affect the user’s ability to achieve their goal​.

5. Google gets stricter on exaggerated or mildly misleading claims

Google’s Search Quality Rater Guidelines ]now explicitly target exaggerated or mildly misleading claims about the creator of a webpage, even if those claims don’t rise to the level of outright deception.

Newly added Section 5.6 explains:

“Deceptive information about a website or content creator is a strong reason for the Lowest rating.”

But it also warns that less blatant exaggerations (e.g., inflated credentials, manufactured expertise) are now enough to warrant a Low rating:

“Sometimes the information about the website or content provider seems exaggerated or mildly misleading, such as claims of personal experience or expertise that seem overstated or included just to impress website visitors.”

This means raters are supposed to rely on what the main content actually demonstrates, plus outside research, rather than taking claims at face value:

“E-E-A-T assessments should be based on the MC itself, the information you find during reputation research, verifiable credentials, etc., not just website or content creator claims of ‘I’m an expert!’”

If a rater finds that the creator’s claimed qualifications feel more like marketing spin than substance, the document is clear:

“If you find the information about the website or the content creator to be exaggerated or mildly misleading, the Low rating should be used.”

Some other smaller changes

Google also made a few other minor changes.

• Lowest quality pages (Section 4.0): Google added this line: “The Lowest rating is required if the page is created to benefit the owner of the website (e.g. to make money) with very little or no attempt to benefit website visitors or otherwise serve a beneficial purpose.”
• Deceptive Page Purpose, Deceptive Information about the Website, Deceptive Design (Section 4.5.3): Google revised this section and added more information, breaking these out with a table and bulleted list with examples.
• New rating type: Low Recipe 3: This will be given to recipe pages with a prominent amount of unrelated content, interstitials, and ads.
• Ad Blocking Extensions (Section 0.4): Raters must now “turn off any ad blocker capabilities of the browser you use to view webpages for rating tasks.” This applies to browsers like Chrome that automatically block some ads.