Back to all articlesContent

Google does not penalise AI content. It penalises useless content

The distinction is not semantic: it completely changes where you invest editorial effort.

Talent Warehouse··2 min read
Orange risograph illustration of a magnifying glass over a messy pile of printed sheets

What the guideline actually says

Google policy on automatically generated content does not mention the tool. It mentions the purpose: content produced at scale with the primary goal of manipulating rankings, without adding original value. An article written with AI, with editorial judgement, verified data and a point of view, complies. An article typed by hand by an intern paraphrasing the top three results does not.

Anyone who kept only the headline that AI gets penalised is optimising the wrong variable.

What the algorithm is really sanctioning

  • Redundancy. Text repeating what already sits on page one, with different synonyms. This is what gets mass-produced and what falls first.
  • Absence of experience. Claims with no source, no case, no first-party number. If your article could have been written by someone who never touched the subject, it shows.
  • Disproportionate scale. A site going from 12 articles a month to 400 with no change in team or structure is an obvious signal.
  • Intent mismatch. Answering with a long essay a query that wanted one specific fact.

The point almost nobody considers

The problem is no longer only ranking in Google. It is being cited by generative engines. Those systems favour material with first-party data, specific figures and attributable claims, because they need something to cite. Generic mass-produced content is exactly what a model can generate on its own, and therefore will never cite.

Put differently: mass AI content production is not only risky for SEO, it is strategically useless for AI visibility. The material that wins on both fronts is the same material.

How we run it

We use AI for what it is good at: structure, first drafts, translation, headline variants, consistency checks. Human input goes where it creates difference: the thesis of the piece, first-party data, examples from real clients, and the judgement to cut what adds nothing.

The publishing rule is a single question: what does this text contain that did not exist on the internet before. If the answer is nothing, it does not ship, regardless of who or what wrote it.

#seo#content#ai#google#strategy