TechCrunch: Amazon shows AI-generated product images in search results

Amazon announced that its shopping app will display AI-generated images based on what users search for. When a customer types a query such as “blue gingham dress,” the app shows synthetic images representing what that item could look like, positioned below the autocomplete suggestions. The images are not linked to any specific listing and are intended to help users refine their search or browse stylistic variations before seeing actual products.

The feature introduces a trust problem that TechCrunch flagged directly: users who don’t read carefully may interpret the generated images as pictures of products that exist in the catalog, then be disappointed when the actual inventory doesn’t match. This expectation gap between what a user sees and what is purchasable is one of the more concrete UX risks in applying generative images to commerce contexts.

Other platforms have navigated this in different ways. Pinterest separates inspiration browsing from product discovery. Google Shopping has tested similar visual refinement features while being explicit about when images are generated. The distinction matters because user intent at the moment of a product search is usually purchase-oriented, not exploration-oriented—which changes the calculus for how misleading an AI-generated image can be before it costs a conversion.

For product managers in commerce and marketplace products, the Amazon experiment is worth tracking as a live test of how generated visuals affect conversion rather than just engagement. The metric that will determine whether this expands is probably click-to-purchase on queries where AI images appeared compared to those where they didn’t. If users click more but complete fewer purchases, the feature is trading trust for engagement. Amazon has not published performance data at this stage.