New Research Sheds Light on Google’s Treatment of AI-Generated Content

The increasing integration of Artificial Intelligence (AI) into content creation has sparked widespread debate and concern within the digital marketing and SEO communities regarding how search engines, particularly Google, evaluate and rank AI-generated material. While AI offers unprecedented potential for automating and augmenting content production, a growing number of websites have experienced significant traffic declines after scaling their output with AI, leading to the phenomenon dubbed "Mount AI" traffic graphs. This has prompted many marketers to shy away from generative AI entirely, fearing punitive measures from Google. However, new research suggests that Google’s stance is not inherently against AI content itself, but rather against low-quality, spammy content, which often overlaps with AI-generated material.

A comprehensive study, leveraging Ahrefs’ AI content detector, aimed to answer critical questions about the relationship between AI-generated content and search engine performance. The research analyzed one million pages from the top 10 positions across 100,000 search engine results pages (SERPs) in June 2026. The AI content detector estimates the percentage of AI-generated text on a page, with results categorized into low (<20%), moderate (20-50%), high (50-80%), and very high (≥80%). Detection was limited to pages with at least 350 words to ensure sufficient text for analysis. It is important to note that AI detection is probabilistic and may differ from Google’s internal methods. The study emphasizes analyzing large-scale datasets and directional trends rather than fixating on individual articles.

AI Content in Top-Ranking Pages

The study found that the majority of top-ranking pages are predominantly human-written. Out of the approximately 150,000 pages analyzed for AI content, 54.7% exhibited less than 20% AI-generated text. Furthermore, pages with under 50% AI content constituted 82.2% of the top-three rankings. While higher AI usage generally correlated with lower ranking positions, the research did reveal that fully AI-generated content can indeed achieve top-three rankings. A notable 9% of top-ranking pages were estimated to be 80% or more AI-generated, with 5.3% being 100% AI-generated. This indicates that while human-written content dominates, AI-generated pages are not entirely absent from high-ranking positions.

Prevalence of AI-Generated Content Across SERPs

Further analysis of AI content distribution across the top 10 SERP positions revealed a trend: highly AI-generated content becomes progressively rarer as one moves up the rankings. Position 1 showed the highest proportion of pages with low AI content and the lowest proportion of pages with very high AI content. However, a gradual gradient was observed, with the prevalence of AI-generated content decreasing incrementally from position 10 to position 1. Even at the top positions, a significant share of pages contained substantial AI-generated text, with 8-12% of pages in each ranking position estimated to be 80% or more AI-generated.

When examining the average and median AI content levels across 15,000 pages per position, a similar pattern emerged. AI content became more common further down the SERP, but still appeared in higher positions. For instance, the average AI content level for position 1 was 27.1%, with a median of 17.1%. As positions descended, these averages slightly increased, reaching 30.9% for position 10 with a median of 19.5%. This data reinforces the notion of a correlation between higher AI content and lower SERP positions, but not an absolute exclusion.

Google’s Indexation of AI-Generated Content

The research also investigated whether AI use impacts a page’s likelihood of being indexed by Google. By analyzing 100,000 pages from a crawler database (one page per domain, with sufficient content for AI detection), the study found that higher AI content levels correlated with lower indexation rates. Pages with very high AI content (≥80%) had an indexation rate of 27%, compared to 49% for pages with low AI content (<20%). However, a substantial 40% of pages with high or very high AI content levels were still indexed, suggesting that Google does not automatically de-index content solely based on its AI origin.

The indexation rates for different AI content buckets were as follows:

• Low (<20% AI): 49% indexation rate
• Moderate (20-50% AI): 39% indexation rate
• High (50-80% AI): 31% indexation rate
• Very High (≥80% AI): 27% indexation rate

This indicates a clear trend of decreasing indexation with increasing AI content, but a significant portion of AI-generated content still makes it into Google’s index.

Long-Term Organic Performance of AI-Generated Pages

To assess the long-term impact on organic performance, the study examined impression trends over 3-6 month periods using Google Search Console (GSC) data. Two panels of websites were analyzed: one from June 2025 to January 2026, and another from January 2026 to June 2026. The URLs were categorized by their estimated AI content level.

While the data from this limited timeframe did not yield definitive conclusions about automatic suppression, several observations were made. Pages with higher AI content levels generally exhibited a slight downward trend in impressions over time. However, this trend was not dramatic, and impressions remained relatively stable for most AI content buckets. The researchers cautioned against assuming automatic suppression, highlighting that many factors influence organic performance. Sites heavily reliant on AI might be newer or have lower authority, and conversely, well-performing sites might be less inclined to publish AI content.

The Crucial Role of Content Quality

The study’s overarching conclusion is that Google is not inherently against AI-generated content but is against low-quality content. The correlation between AI use and lower performance—reduced impressions, lower ranking positions, and lower indexation rates—is likely due to the fact that AI-generated content frequently overlaps with characteristics of poor-quality content. These common issues include:

• Lack of depth and originality: AI can sometimes produce superficial or repetitive information.
• Factual inaccuracies: While improving, AI can still generate incorrect information.
• Poor readability and unnatural phrasing: AI-generated text may sometimes lack human-like flow and clarity.
• Generic and unengaging content: AI may struggle to capture unique voice or emotional resonance.
• Lack of unique expertise or experience: Content may not reflect firsthand knowledge or deep insights.

These quality deficiencies, which are often exacerbated when AI content is produced at scale, can significantly harm organic performance. As explained by Dan Taylor in a recent Search Engine Journal article, publishing low-quality, AI-generated content at scale can lead to a "Mount AI" situation where Google initially crawls these pages but eventually removes them from the index due to a lack of sustained quality and authority, a phenomenon referred to as "burst-crawling."

Expert Takeaways

The research indicates that while AI use is correlated with lower search performance, this is not a hard cutoff. AI-generated content is present in all ranking positions, and a significant portion is successfully indexed. The gradual performance decline with increasing AI use is likely attributable to a general decrease in content quality, rather than an explicit penalty for AI authorship.

Marketers and SEO professionals are advised not to fear AI content creation itself, but rather the pitfalls of creating spammy and low-quality content. The study suggests that it is possible to create detailed, interesting, and well-researched content using AI, but this requires careful oversight and editing to ensure high quality. The key takeaway is that Google continues to prioritize content quality, regardless of its origin. The challenge lies in ensuring that AI-assisted content meets these established quality standards, rather than falling into the trap of mass-producing low-value material.