Etsy now requires one real photo per listing: why to judge your AI visuals on returns, not just conversion
Alexandre Suon·8 min read·
Since October 22, every Etsy listing must show at least one real, unedited photo of the item, and bot-generated listings are banned. What the new rules say, why the marketplace is protecting trust, and what to test on your product pages.
In this articleShow all sectionsHide
30-second summary
ShowHide ⌄
- Most Etsy listings must now include at least one unedited photo of the actual item, against a real background.
- Other images can still be edited, including with AI, as long as they accurately represent the item.
- Bot-generated listings without human input and oversight are banned and will be removed.
- Buyers can report misleading AI imagery, and may get a refund if the listing did not meet the image rules.
- For a brand, a visual that lifts conversion can also lift returns: that is the number to watch.
Section 1 · The news
Etsy requires at least one real photo per listing
On October 22, 2026, Etsy updated its Seller Handbook with new rules on AI use. The most visible change: most listings must show at least one unedited photo of the actual item, taken against a real background.
Etsy goes further. Listings generated by bots or mass-automated without individual human creative input and oversight are prohibited. Buyers can flag AI imagery they find misleading. And the marketplace says it is investing in tools to detect and remove content that breaks these rules.
AI is not banned outright. Etsy frames the rules as a first step: James Ossman, VP of Customer Operations, calls them "our starting point, not the finish line."
Section 2 · The rules
What is still allowed, and what no longer is
| Topic | What Etsy says |
|---|---|
| Real photo | At least one photo of the actual item, on a real background, for most listings |
| Edits to that photo | Not allowed: no trimming, scaling, leveling, lighting or exposure changes, or background replacement |
| Other images | Editing and AI allowed, if the image accurately represents the item |
| AI disclosure | Recommended for AI editing, required when AI creates the product itself |
| Patterns, recipes, tutorials | Can no longer be AI-generated, must show real results |
| Automated listings | Banned without human input and oversight, and removed |
Why a single photo changes the game The rule does not ask sellers to give up edited visuals. It sets a reference point: the buyer sees the item as it really is at least once. If the other images drift too far from it, the gap becomes visible on the listing itself.
Section 3 · Enforcement
Detection, reports and refunds
- 1At uploadThe listing form flags images detected as possibly AI-generated.
- 2Through buyersBuyers can report misleading AI imagery. At first, reports are mainly used to gather information.
- 3After purchaseA buyer who opens a "materially different" case may be refunded if the listing did not meet the image rules.
- 4For the shopRemovals will expand over the coming months. Repeat violations can lead to shop removal.
The link to refunds is the key point for sellers: an overly flattering image is no longer just a risk of a bad review, it is grounds for a refund under the marketplace's own rules. A penalized seller can appeal from the Policy Violations page in Shop Manager.
Section 4 · The trap
A visual that looks better than the product lifts conversion, then returns
Many brands now produce part of their product visuals with AI: it is cheaper, faster and often more flattering. The risk is not AI itself. It is judging those visuals on a single number, the conversion rate.
- 1The visual wins attentionMore clicks, more add-to-carts: the test looks like a win.
- 2The parcel arrivesThe color, material or size does not match what the customer saw.
- 3The customer sends it backReturn, refund, shipping costs, sometimes a negative review.
- 4Margin dropsOnce returns are counted, the winning visual can earn less than the old one.
“The visual that wins on conversion is not always the one that wins on margin once returns are counted.”
Section 5 · What to test
Test AI visuals like any other product page change
Etsy's rule only applies to Etsy. But the question it raises applies to any online store: do your visuals show what the customer will receive? The only way to know is to measure after delivery, not just at the moment of the click.
- Test AI visuals against real photos, category by category, with a control group, rather than switching everything at once.
- Pick the metric before launching the test: conversion, but also return rate, return reasons and review scores.
- Wait for the return window to close before calling a winner: a few extra weeks avoid a bad decision.
- Keep a real photo clearly visible where looks matter: fashion, materials, colors, home decor.
- Read return reasons such as "does not match the photo": it is the most direct signal of a misleading visual.
Section 6 · Open questions
What the announcement does not specify yet
Etsy has not given a compliance deadline or a grace period for existing listings. The rule covers "most" listings, with no detailed list of exceptions in the sources reviewed. And the level of human involvement needed for a listing not to count as automated is not quantified.
Etsy also says it will explore more ways to tell buyers when AI has been used. More announcements are likely to follow.
Key terms
ShowHide ⌄
- Unedited photo
- For Etsy: a photo of the actual item, on a real background, with no trimming, scaling, lighting changes or background replacement.
- Automated listing
- A listing created by a bot or in bulk, without human input and oversight. Etsy now prohibits it.
- "Materially different" case
- A buyer claim that the item received does not match the listing. It can lead to a refund.
- Return rate
- The share of orders sent back. It is measured after delivery, so several weeks after the conversion.
- Control group
- A share of visitors who keep the old version. Comparing them with the rest measures the real effect.
Sources
ShowHide ⌄
Alexandre co-founded and runs Henkan & Partners. He decodes one e-commerce story every day on LinkedIn.
Follow Alexandre ↗
