The study that's making content teams nervous
Semrush published research in April 2026 that every marketing team running an AI content operation should read. As reported by Search Engine Land and in Semrush's own study, the firm analysed 42,000 blog pages across 20,000 keywords and used the GPTZero detector to classify each page as human-written, AI-assisted, or fully AI-generated.
The headline number: human-written content appeared in the number-one position 80% of the time. Purely AI-generated content appeared there 9% of the time.
The part most people aren't talking about
The same study surveyed 224 SEO professionals. According to Search Engine Land's coverage, 72% of them believe AI content performs as well as or better than human-written content — the ranking data says otherwise at the very top of the page. Semrush also found that when you widen the lens to the whole top 10, AI and human content actually reach page one at similar rates (57% vs. 58%); the gap is concentrated at position one specifically.
Semrush and Search Engine Land both flag a real limitation: AI detectors like GPTZero are known to misclassify content, which introduces some noise into these figures. Treat the 8x figure as directionally real, not laboratory-precise.
🔑 Key finding, with the caveat attached
Human-written content is 8x more likely to hold the #1 spot, per Semrush's analysis of 42,000 pages. AI detection is imperfect, so treat the exact multiplier as an estimate — the direction of the finding is the more durable takeaway.
What Google itself says about AI content
Separately from the Semrush data, Google's Search Central documentation is explicit that its ranking systems reward content quality, not the method used to produce it. Using generative AI to add structure or accelerate research is fine; using it to publish "many pages without adding value for users" falls under Google's spam policy on scaled content abuse. Google also recommends giving readers context about how automation was used in creating a page.
Put together, the two sources tell a consistent story: Google doesn't penalise AI as a production method, but AI drafts on average lack the specific signals — original experience, editorial perspective, proprietary data — that both Google's quality systems and human readers reward.
What this means for your AI content workflow
If you're using AI to write first drafts and publishing without meaningful human review, this data is a direct warning. Quality signals AI drafts consistently lack:
- First-person experience or specific examples from actual work
- Original data, proprietary research, or unique case studies
- Editorial perspective — a genuine point of view, not a balanced overview
- Specificity that only comes from doing the thing being written about
Semrush's companion survey found teams overwhelmingly keep a human in the loop: 87% keep humans heavily involved in content creation, and 64% run a human-led, AI-assisted workflow, using AI mostly for research, drafting, and optimisation rather than final judgment calls. That's the same principle behind how BuzzRiding approaches its own posts — and it's the corrective this article itself needed, since an earlier version implied a hands-on comparison this site never ran.