The Data Point
The study analyzed 20,000 keywords and 42,000 blog posts, finding that content classified as purely AI-generated appeared in the top spot just 9% of the time, while human-written content was there 80% of the time. The study used GPTZero to classify each article as human-written, AI-generated, or mixed, and calculated the probability of each type of content appearing in the top 10 SERP positions.
Why the Algorithm Does This
The mechanism behind this finding is that search algorithms reward human originality and creativity, which is harder to replicate with AI-generated content. The study suggests that AI-generated content may lack the nuance and depth of human-written content, which is why it performs worse in search rankings. This is consistent with the finding that 70% of SEO teams cite speed as the top benefit of using AI, but only 19% say it improves content quality.
The Creator / Developer Play
To take advantage of this finding, creators and developers can use AI to move faster through research, outlining, and drafting, but then invest the time saved into incorporating expert insights, proprietary data, and other elements that make their content stand out. For example, they can use AI to generate a first draft, and then have a human editor review and refine the content to add depth and nuance. They can also use tools like Semrush’s Content Toolkit to generate content briefs that reflect what’s ranking and what’s missing, and then add their human insights and perspectives to take the generated content to the next level.
What the Research Doesn’t Cover
The study has some limitations, including a sample size of 42,000 blog posts, which may not be representative of all types of content. Additionally, the study only analyzed content from November 2025, which may not reflect changes in search algorithms or AI technology since then. Furthermore, the study only used GPTZero to classify content as human-written, AI-generated, or mixed, which may not be the only or best way to detect AI-generated content. These limitations suggest that further research is needed to fully understand the relationship between AI-generated content and search rankings.