11
New Research Sheds Light on Google’s Treatment of AI-Generated Content, Highlighting Quality Over Origin
New research analyzing millions of web pages suggests that Google does not inherently penalize AI-generated content. Instead, the search engine continues to prioritize content quality, with a significant overlap between AI-generated material and content that is considered low-quality or spammy. This nuanced finding challenges widespread fears among marketers that the use of generative AI will automatically lead to search engine penalties.
Automation has been a long-standing practice in content creation, with AI offering new avenues for assistance and generation. However, this has also led to concerns, as noted by SEO experts Lily Ray and Glenn Gabe, who have documented instances of websites experiencing dramatic spikes in traffic and rankings after scaling content creation with AI, only to see their performance plummet months later. These "Mount AI" traffic graphs illustrate the potential risks associated with this approach.
The core question for many marketers is whether Google is punishing the use of AI itself, or if the penalties are a consequence of the type of content often produced at scale with AI – namely, spammy or low-quality content. To address this, a comprehensive study was conducted to answer key questions about AI content and its performance in Google search results.
Methodology: A Probabilistic Approach to AI Detection
The research employed Ahrefs’ AI content detector, a tool accessible to Ahrefs customers through Site Audit and various Site Explorer reports. This detector estimates the percentage of AI-generated text on a page, categorizing it from low (<20%) to very high (≥80%). For accuracy, detection is only triggered on pages exceeding 350 words.
It is crucial to acknowledge that AI content detection is not infallible and likely differs from Google’s internal methods. Similar to how Large Language Models (LLMs) use probability to generate text, AI detectors use probability to classify it. The study’s authors emphasize that the most valuable insights come from analyzing large-scale datasets and directional trends, rather than fixating on individual articles and their specific AI content scores.
Key Findings: AI Content in Top-Ranking Pages
The study analyzed 1,000,000 pages from the top 10 positions across 100,000 search engine results pages (SERPs) in June 2026. Of these, approximately 300,000 were available in the crawler database, with 150,000 containing sufficient content for AI detection.
The majority of top-ranking pages were found to be predominantly human-written. A substantial 54.7% of these pages contained less than 20% AI content, and pages with under 50% AI content accounted for 82.2% of top-three rankings.
Despite this trend, the research indicates that fully AI-generated content can indeed rank in the top three positions. While a small minority, 9% of top-ranking pages were estimated to be ≥80% AI content, and 5.3% were entirely AI-generated.
AI Content Distribution Across SERP Positions
Further analysis revealed that highly AI-generated content becomes less common as one moves higher up the SERP. Position 1 exhibited the highest proportion of pages with low AI content and the lowest proportion of pages with very high AI content. However, a gradual decline in the prevalence of AI-generated content was observed from position 10 down to position 1, with relatively small increments. Even at position 1, between 8-12% of pages contained ≥80% AI content.
Breakdowns of AI content levels across SERP positions illustrate this gradient:
| Position | Very High (≥80%) | High (50-80%) | Moderate (20-50%) | Low (<20%) |
|---|---|---|---|---|
| 1 | 8.4% | 8.1% | 27.5% | 55.9% |
| 2 | 9.0% | 8.8% | 27.6% | 54.6% |
| 3 | 9.6% | 9.5% | 27.4% | 53.5% |
| 4 | 9.8% | 10.0% | 26.9% | 53.3% |
| 5 | 10.0% | 10.1% | 26.9% | 53.0% |
| 6 | 10.7% | 10.2% | 26.5% | 52.6% |
| 7 | 10.2% | 10.1% | 27.7% | 52.0% |
| 8 | 11.1% | 9.8% | 28.0% | 51.0% |
| 9 | 11.3% | 10.2% | 28.1% | 50.3% |
| 10 | 11.7% | 9.5% | 27.7% | 51.1% |
An examination of average and median AI content levels across 15,000 pages per position further supported this observation, showing a similar trend of AI content becoming more common further down the SERP, yet still appearing in higher positions.
| SERP Position | Average AI Content Level | Median AI Content Level |
|---|---|---|
| 1 | 27.1% | 17.1% |
| 2 | 28.1% | 17.7% |
| 3 | 28.9% | 18.2% |
| 4 | 29.4% | 18.3% |
| 5 | 29.6% | 18.3% |
| 6 | 30.1% | 18.6% |
| 7 | 29.8% | 18.9% |
| 8 | 30.7% | 19.4% |
| 9 | 31.2% | 19.9% |
| 10 | 30.9% | 19.5% |
Indexation Rates and AI Content
The research also investigated whether Google indexes content with a high likelihood of being AI-generated. By analyzing 1,000,000 pages from the crawler database (one page per domain) and filtering for those with sufficient content for AI detection, the study found that higher AI content levels correlate with lower indexation rates.
However, a significant portion of pages with high or very high AI content levels (40%) were still indexed, a rate not drastically lower than the 49% indexation rate for low AI content pages. This suggests that while AI use might slightly reduce the chance of indexing, it does not guarantee exclusion.
Long-Term Traffic Trends for AI-Generated Content
Finally, the study examined the organic performance of AI-generated pages over time, comparing them to non-AI-generated pages. Using Google Search Console (GSC) data from live URLs and matching them with the crawler database, pages were categorized by their estimated AI content level.
The analysis of impression changes over time, across two distinct panels of websites, showed that organic performance is influenced by numerous factors. While the data from this limited timeframe does not allow for definitive conclusions, it is noted that:
- There is no clear evidence of a significant decline in impressions for AI-generated content over 3-6 months.
- The organic performance of AI-generated content remained relatively stable over the observed periods.
- The study did not find concrete evidence of automatic suppression of AI-generated content.
The authors caution against assuming automatic suppression, highlighting that sites heavily reliant on AI content may also be newer or have lower authority. Conversely, sites with strong organic performance might be less inclined to publish AI content. The most probable explanation, according to the research, is that an increasing use of AI correlates with a decrease in content quality.
Conclusion: Quality Remains Paramount
The research concludes that AI use appears to be correlated with lower search performance metrics, including fewer impressions, lower ranking positions, and reduced indexation rates. However, this correlation does not equate to causation or a direct penalty. Significant amounts of AI-generated content are successfully indexed and rank highly, and their impressions remain relatively stable over time.
The data does not suggest hard cutoffs indicating a binary AI classifier preventing ranking or indexing. Instead, a gradual gradient exists where performance generally worsens with increased AI use, likely because content quality also tends to decrease with higher AI generation.
The study reiterates that when companies adopt AI content strategies, they often engage in practices that differ from traditional content creation and SEO, frequently resulting in lower quality. Common issues with basic AI content include:
- Lack of depth and originality
- Repetitive phrasing
- Factual inaccuracies
- Poor readability
- Lack of unique insights or expertise
These issues, which can negatively impact any content, are particularly prevalent in AI-generated material. Publishing such content at scale can lead to the "Mount AI" phenomenon, where Google initially crawls new pages but may remove them from the index if the domain lacks the authority to sustain that volume of low-quality content.
Ultimately, the research suggests that Google is not actively punishing AI-generated content but rather relying on its established signals of content quality. The challenge lies in the fact that AI-generated content frequently exhibits lower quality than human-generated content. While quality problems commonly co-occur with AI use, they are not an inherent guarantee. The study acknowledges that creating detailed, well-researched content using AI is possible, though it is not yet the common practice. Marketers and SEO professionals are advised to focus on avoiding spam and scaled, low-quality content rather than fearing AI content creation itself.
