Etsy sellers can still offer some AI-assisted artwork, but they cannot sell functional instructions generated by AI. The platform's October 22, 2026 update also tightens what listing images must demonstrate. For a knitting pattern or recipe, changing the disclosure alone will not answer the central questions: who wrote the instructions, and does the pictured result exist?
Etsy announced the changes on October 22; The Verge reported them on October 23 in UTC. The announcement and the updated policies share the October 22 date. Sellers should review existing listings as well as new uploads.
Instructions must stand up as the seller's own work
Etsy's AI-creations guidance distinguishes an expressive work from information that directs a buyer through a task. AI-generated craft patterns, recipes, project plans and instructional guides fall into the prohibited functional category. Seller-prompted artwork remains permitted when the seller makes an individual creative contribution and discloses the use of AI.
That distinction matters for digital products. A convincing cover or a polished preview does not establish that the steps inside a guide are usable. A seller reviewing a pattern should examine its instructional content separately from its promotional images. Having personally produced a finished object is useful evidence about the pattern, but it does not turn AI-written instructions into permitted content.
Nor is a general statement that a shop uses AI a substitute for explaining its use in the relevant item's description. Etsy's policy concerns the individual creation, rather than a blanket permission to automate a catalogue.
The image rule follows the product type
The updated Listing Image Requirements generally require an unedited photograph of the actual item against a real background. Other pictures may use editing tools, including AI, provided they accurately represent what the buyer will receive. Removing a flaw or changing a material's appearance would defeat that purpose.
The exceptions deserve attention. Authorized stock mockups can be used for qualifying print-on-demand items; many digital downloads need a preview or excerpt; custom and made-to-order work can use photographs of real finished examples. Patterns and similar products need a real, unedited photograph of a result made using the product.
A useful listing review therefore starts with the category. Identify which image is the evidence of the item or outcome, then assess whether the remaining illustrations create different expectations. A beautiful simulated result cannot do the job of the required real example.
Enforcement tools are still developing
Etsy says it is introducing image guidance during listing creation and stronger detection. Its new reporting reason for potentially misleading AI imagery will initially help it understand reports, rather than trigger immediate enforcement.
Those plans do not show how reliably the detection works. Buyers should still read the description and inspect the actual-item evidence; sellers should correct a misleading listing without waiting for a warning. Etsy says existing listings must meet the updated image requirements and that a materially different delivered item can lead to a refund for which the seller is responsible.
Sources
Related Articles
vLLM TorchTPU reuses PyTorch models, with validation tied to each TPU
The new backend keeps vLLM’s model definitions and serving interface. Hardware coverage, checkpoint validation and compilation still need separate evaluation.
RTX Spark brings CUDA PyTorch to Windows Arm64, with dependency gaps
Nvidia documents native Windows Arm64 builds as RTX Spark laptops arrive. Developers still need to match Python wheels, extensions and drivers to the complete workload.
Microsoft exits Nscale's Loughton plan as power delivery remains unresolved
The withdrawal concerns one planned UK data centre. Company responses preserve the wider partnership, while the original GPU target still needs a credible route to power.
VitVio's Vetro rollout in Gdańsk puts operating-room timelines first
The selected-room deployment targets surgical-stage logging and coordination. Its value depends on measured workflow results and clear handling of captured data.