There’s a persistent fear in SEO circles that Google algorithmically detects and demotes AI-written content. A new study from Ahrefs, covering 331,000 pages, challenges that assumption.
What the Data Shows
Google’s own public position has been consistent: they don’t care whether AI was used to produce content, provided that content wasn’t built primarily to manipulate search rankings. The Ahrefs study, authored by Ryan Law, puts data behind that claim.
The finding is that Google penalizes bad content, not AI content. The distinction matters for how you think about your publishing workflow.
The Operator Implication
If you’re producing AI-assisted content that’s genuinely useful, well-researched, and written for a real reader, the signal this study sends is that Google’s systems aren’t targeting AI origin as a ranking factor. What they’re targeting is thin, manipulative, or low-effort content, regardless of how it was produced.
The reverse is also true. Human-written content that’s padded, keyword-stuffed, or designed to rank rather than inform is just as exposed as its AI equivalent.
Why This Matters Now
Debates about AI content detection tools and “AI penalties” have pushed some operators toward over-correcting: manually rewriting AI drafts purely to obscure AI origin, adding friction without improving quality. This study suggests that energy is better spent on the content itself than on disguising how it was made.
The full study is on the Ahrefs blog. Worth reading if you’re making decisions about AI content volume and editorial standards in 2025.
