Every seller knows the drill: a product photo snapped on a wrinkled bedsheet or a cluttered kitchen table kills your click-through rate (CTR) in search before a buyer even reads the title. To save sellers the cost of a studio shoot, the eBay mobile app rolled out a built-in diffusion AI tool. It automatically cuts the item out of the frame and drops it into a realistic setting—neutral studio lighting, marble, wood grain, or an interior backdrop. On desktop, you also get a one-click swap to a classic white background.
Alongside the generator, however, the marketplace tightened its moderation rules (eBay Picture Policy). A polished cover shot cannot cross the line into misleading the buyer:
- No hiding flaws: the AI is strictly barred from 'smoothing out' scuffs, chips, or scratches on the item itself.
- No stock renders: for used and refurbished goods, you cannot replace real photos of your actual item with generic catalog shots.
- Show wear up close: any signs of use must appear clearly in additional unedited photos.
If a generated background distorts the item's true color or masks damage, the buyer can open an Item Not As Described (INAD) dispute. eBay will deduct return shipping straight from your balance, and the Cassini search algorithm will demote your listing in search results.
Just like we saw with mandatory labeling for synthetic models on Amazon, global e-commerce draws a hard line between marketing polish and honesty: use AI to build a clickable main cover, but always back it up with raw, accurate photos on the remaining slides.
