The short answer
Yes, and the size of the discount depends almost entirely on whether the traffic held. Buyers are not making an aesthetic judgement about how your first draft was written. They are using AI content as a signal for something they cannot otherwise measure quickly: how likely this site is to lose rankings in the next twelve months. A site with unedited AI content and stable two-year traffic is treated very differently from one with the same content and six months of history.
What the published data claims, and who published it
The figures circulating are that human-written sites sell at a 39% higher profit multiple than AI-heavy ones, and that AI-content sites take 54 days to sell against 35 — around 54% longer. Both come from an analysis of Motion Invest listing data published by Originality.AI. Worth knowing before you lean on them: Motion Invest is a marketplace that sells websites, and Originality.AI sells AI-detection software, so both parties have an interest in the finding. I have not been able to verify the numbers against the primary study. Treat the direction as well supported and the precise percentages as one marketplace's data reported by an interested party.
Why buyers discount it at all
Because content produced at volume without editing has a track record of losing rankings, and a buyer cannot easily tell from the outside whether yours is the exception. The concern is not that a machine wrote it. It is that a site which published a hundred pages in a month has demonstrated a process that search engines have repeatedly devalued, and the buyer would be inheriting that process along with the traffic. Rewrite the same content by hand and the concern disappears, which tells you the objection was always about durability.
How they find out
Working due diligence checklists include a content read for exactly this. The usual method is not a detection tool but a sniff test: read four or five articles and see whether they sound like a person who knows the subject. Generic structure, confident vagueness, no first-hand detail, the same three-paragraph shape on every page, and advice that could apply to any niche are what give it away. Detection tools get used too, and they produce false positives, which is a reason to be able to explain your process rather than argue with a score.
AI-assisted and AI-published are different things
This is the distinction that decides which side of the discount you are on. A site where AI produced drafts that a knowledgeable person then restructured, fact-checked and added original material to reads as edited content, because that is what it is. A site where prompts became posts reads as what it is too. Buyers in practice care about the second and are largely indifferent to the first, and the published data on multiples is describing sites at the unedited end rather than every site that ever used a tool.
What to do if you want to sell and have AI content
Rank your pages by traffic and fix the top of the list rather than everything. Add something to each that a model could not have produced: a real test, a number you gathered, a photograph you took, a specific example with a name and a date. A content refresh on the pages that still earn beats rewriting the ones that never did, so consolidate the thin pages rather than reworking them, since a smaller site with nothing embarrassing in it prices better than a large one a buyer has to audit. Two quarters of this work in front of a sale is usually worth more than the same effort spent adding pages.
Whether to disclose it
Yes, and describe the process rather than answering the question narrowly. A seller who says the drafts were AI-assisted and then edited by somebody with subject knowledge, and can show what that editing looked like, has given the buyer a way to price it. A seller who says no and is contradicted by the second article the buyer reads has ended the deal, and not over the content. This is the same logic as disclosing a traffic decline: the discovery is certain, so the only variable is who framed it.
What actually protects the multiple
Traffic that held through several core updates, revenue that does not depend on a single programme, an audience that arrives by means other than search, and a documented process a new owner can continue. Those four are what a buyer is really buying, and a site that has them will survive an AI-content conversation. A site that lacks them will not be rescued by having written everything by hand, which is the part of this debate that gets lost.
- The discount tracks traffic risk, not authorship. Stable traffic through core updates neutralises most of it.
- The 39% multiple gap and 54-versus-35-day figures come from a marketplace’s data reported by an AI-detection vendor — direction reliable, precision less so.
- Buyers find AI content by reading four articles, not by running a detector.
- AI-assisted and AI-published are treated completely differently.
- Fix your top-traffic pages and consolidate the thin ones; do not rewrite everything.
Work it out: Profit multiple calculator
Run a free appraisal to see what your site is worth on its current earnings, and which factors are moving your multiple up or down.
Value my site →