Google is now guiding its quality raters to specifically watch for web pages where the main content (MC) is created using automated or generative AI tools — and to classify such pages as “Lowest Quality”, according to John Mueller, Whether Content Is AI-Generated? Google’s Senior Search Analyst and Search Relations Team Lead. Mueller shared this insight during Search Central Live in Madrid, as posted by Aleyda Solis on LinkedIn.
This directive is part of the January 2025 update to Google’s Search Quality Rater Guidelines. Here’s a breakdown of the most notable changes in that update:
1. Generative AI Officially Defined
For the first time, Google has included a formal definition of generative AI in its guidelines (Section 2.1):
“Generative AI is a type of machine learning model that uses learned data to generate new content — such as text, images, music, or code. While it can be a helpful tool, it can also be misused.”
This addition clarifies Google’s stance: generative AI is useful but potentially problematic, especially when misused for large-scale content creation.
2. Major Overhaul of Spam Definitions
Google has restructured and expanded how it classifies spam, with added emphasis on AI-driven content. Key updates include:
- Expired Domain Abuse (4.6.3): Buying expired domains and populating them with low-value content just to boost rankings.
- Site Reputation Abuse (4.6.4): Hosting low-quality third-party content to exploit a high-authority domain’s reputation.
- Scaled Content Abuse (4.6.5): Producing massive amounts of content with little originality or manual input — often using AI.
- Low-Effort, Low-Originality Main Content (4.6.6): Pages primarily composed of copied, paraphrased, or AI-generated content with no added value.
This last section is what John Mueller emphasized. It states:
“The Lowest rating applies if all or nearly all MC on the page — whether text, images, video, etc. — is copied, paraphrased, embedded, auto-generated, or AI-generated with little to no originality or value for users. Even if credit is given, these pages should be rated Lowest.”
3. Spotting AI or Auto-Generated Content
While the guidelines don’t offer a foolproof method for identifying AI-generated content, they provide new clues for spotting paraphrased content, which is often AI-driven:
- Section 4.6.6: Automated tools may rephrase or summarize content from other sources.
- Section 4.6.7: Paraphrased content typically:
- Contains only basic, widely-known info.
- Closely resembles content from sources like Wikipedia.
- Summarizes pages (e.g., news or forums) without contributing new insights.
- May include telltale language like “As an AI language model…”
4. Clarifying Low vs. Lowest Ratings
Google has drawn a clearer line between “Low” and “Lowest” quality ratings:
- Low: Some effort or curation is present, even if content is reused.
- Lowest: Most or all content is copied, paraphrased, or auto-generated with no added value.
Examples of content likely to receive these ratings:
- Reposts from social media with minimal commentary.
- Pages with embedded videos or “repinned” images and no original input.
- “Best of” lists that rehash existing reviews with little original research.
5. Filler Content Now Penalized
A new section addresses “filler” content — content that inflates a page’s appearance without adding real value:
“Filler can make a page look rich while actually hiding or diluting helpful content.”
This includes:
- Excessive ads or bloated introductions.
- Unnecessary repetition or verbose paragraphs.
- Layouts that bury useful content beneath visual clutter.
Raters are instructed to evaluate whether the layout and hierarchy help or hinder users in achieving their goals.
6. Tighter Rules on Misleading Claims
Google now explicitly warns against exaggerated or mildly misleading statements about a website or content creator (Section 5.6). Even if it doesn’t rise to full-on deception, it can earn a Low rating.
“If you find the information about the website or content creator to be exaggerated or misleading, the Low rating should be used.”
Claims like inflated credentials or vague references to expertise must be backed by real evidence in the main content or through external reputation research.
Raters are urged not to accept self-claims at face value — credibility must be earned, not declared.
Other Noteworthy Changes:
- Lowest Quality Pages (4.0): Now explicitly includes pages created solely to benefit the website owner (e.g., monetization) without serving the user.
- Deceptive Design/Intent (4.5.3): Expanded with examples and clearer formatting.
- New Rating Type – Low Recipe 3: For recipe pages cluttered with irrelevant content and ads.
- Ad Blockers (0.4): Raters must now disable any ad-blocking tools when assessing pages.
Final Thoughts
These updates show Google is taking a harder stance on low-effort, AI-generated, and deceptive content. Quality raters are now instructed to penalize pages that lack originality, over-rely on automation, or attempt to manipulate users with filler or exaggerated claims.
