AI-generated content is text, images, audio or video produced wholly or partly by a generative AI tool, such as ChatGPT, Gemini or an image generator, rather than written or made entirely by a person. In SEO the phrase usually means web pages and blog posts drafted by a large language model.
How Google treats AI-generated content
Google’s stated position is that it rewards helpful, reliable content however it was produced, and that using automation mainly to manipulate rankings breaks its spam policies. In 2024 it introduced a scaled content abuse policy aimed at mass-producing pages with little value, whether by AI, by people or by both. Google’s quality rater guidelines were updated in 2025 to tell raters that pages whose main content is auto-generated with little effort, originality or added value deserve the lowest rating.
So the useful question is not “was this written by AI?” but “does this page help someone, and was it produced at scale to catch search traffic?” A carefully edited article that started as an AI draft can be perfectly fine. Three hundred near-identical town pages generated in an afternoon are not.
Why it matters
Because AI makes text cheap, the web is filling with pages that say the same things in similar words. Content that only restates what already ranks adds no information gain, so it has little reason to be shown or cited. What stands out is what a tool cannot invent: your prices, your process, your photographs, your mistakes and what you learned from them.
Accuracy is the other risk. Language models produce fluent text that can be wrong, a failure called hallucination. In UK marketing that might mean an outdated VAT threshold, a misstated consumer right or a made-up regulation. A false claim is still your claim: the ASA has made clear the CAP Code applies to ads however they were produced. Fake reviews, including AI-written ones, are a banned practice under the Digital Markets, Competition and Consumers Act 2024.
Images raise similar questions. AI-generated photos of a team, premises or finished job that never existed mislead customers in the same way as invented text, and stock-style AI images with odd details can make a small business look careless. Use real photographs of your own work wherever the image is meant to show what you do.
Common mistakes
- Publishing raw drafts. Unedited AI text tends to be generic, padded and full of the same phrases as everyone else’s.
- Scaling page counts. Hundreds of location or keyword-variant pages with swapped place names invite a spam action.
- No checking. Figures, dates, legal points and quotations published without verifying them against a primary source.
- Invented experience. First-person stories or case studies that never happened damage trust and may mislead consumers.
- AI-written reviews or testimonials. A review that does not come from a real customer is a fake review, which UK consumer law bans however it was written.
How to act on it
Use AI where it saves time without replacing judgement: outlines, summaries of your own notes, first drafts of routine copy, alternative headlines. Then add what only you can supply, edit for your voice and remove filler. Check every fact against a primary source before publishing; a simple fact-checking step catches most errors.
Put a real author with relevant experience on each piece, which supports E-E-A-T, and review any older AI-assisted pages for thin or outdated material. If you want a content plan built around this way of working, that is part of my content SEO and strategy service.