Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- AI images are allowed when product information is accurate. An AI label does not make a misleading image acceptable.
- The September 24 document applies to the United States. It does not establish same-day enforcement in Malaysia, Vietnam or globally.
- Collect physical-product photos, specifications, SKUs and licensing evidence first, then compare each image with the title, description and attributes.
- Generated effects must not change dimensions, materials, set quantities, accessories or compatibility. Flat products must not appear to have three-dimensional physical structures.
- This article provides a pre-upload checklist, not store test results or penalty case studies. Actual review outcomes and penalties depend on platform rules and the specific facts.
The direct answer: AI is allowed, but must accurately depict what is sold
The US Seller University product-image document is dated September 24, 2026. It allows AI-generated, rendered or edited images that accurately represent the physical product. The problem arises when generated visuals change the buyer's understanding of what is being sold. Sellers, production teams and store-management agencies should review images against the actual delivered item.
The page does not provide a separate future effective date or a uniform penalty count. Its date is not evidence that every store has already been penalized. This page retains the original event's unique URL, but the directly verifiable scope of this review is the United States; it is not automatically extended to Malaysia, Vietnam or other markets.
Prerequisite: Build evidence linking the physical product, PDP and SKU
Collect actual product photos, dimension and material information, a variant list, package quantities and substantiated compatibility descriptions. Group images with titles, descriptions and attributes by product ID, SKU and asset version. If information is missing, do not generate substitute claims about product effects. Separately check licenses when using other people's materials or likenesses.
The US Product Listing Policy requires accurate, consistent product information and prohibits changing the content of already-purchased orders by modifying the detail page. Confirm the selected SKU and delivered item before reviewing images. Do not evade problems by changing titles or categories or creating duplicate listings.
Step one: Check whether the visual changes the product buyers see
Compare appearance, dimensions and structure image by image. Generated shadows must not disguise flat stickers or printed artwork as three-dimensional carvings; scale references must not make objects appear substantially larger; and the featured object must be the product being sold. Clothing scale may be subject to separate category-review standards. Check the category's specific requirements first, and do not treat that exception as permission to change dimensions arbitrarily.
Then check each SKU's quantity, set contents, accessories, materials and intended environment. Do not depict a multipiece set when only one item is delivered. Do not add unsupported installation locations, compatible devices or functions such as illumination. If comparative evidence is missing, remove those claims and pause use of the asset.
Step two: Review AI disclosure and accuracy separately
The US general AIGC rules dated September 2 require clear disclosure for fully generated or substantially AI-edited content. Basic adjustments such as slight brightness changes, cropping or noise reduction should not automatically be treated as substantial alteration. For video, use the label or in-content disclosure available at the actual publishing entry point. Do not infer that every PDP image has the same button as the video setting.
Disclosure describes the production method; it does not remove product-accuracy or licensing requirements. For videos and promotional assets, also check for fabricated identities, endorsements or unsupported effects. Platform-added labels and your own disclosure are separate paths. The absence of an automatic label does not mean disclosure is unnecessary.
Step three: Verify, stop and rework
Use a per-SKU review sheet to record the asset version, physical-product evidence, corresponding PDP fields, how AI was used, licensing and reviewer. In the preview, check that buyers see the correct image and quantity when choosing different SKUs. These are recommended procedures: this article did not sign into a store, upload assets or verify approval.
Stop uploading if specifications conflict, effects are unsupported, source licenses are unclear or the actual entry point cannot provide proper disclosure. Rework the image or use real photos backed by evidence. Save originals, modification records and platform notices. Even restored older assets must be checked again for compliance; withdrawing a change cannot be guaranteed to undo an existing review or penalty.
Impact and next steps: Review under each market's rules
The platform determines action according to the nature, severity and frequency of violations. Without independent store results, do not claim that a first violation guarantees a ban, AI use guarantees reduced reach or a label guarantees approval. If instructed to make corrections, retain evidence according to the notice and current platform procedure; do not repeatedly submit the same questionable asset.
For multimarket content, record the target marketplace, rule version and category separately. Do not use the US document as a global template. If you receive a store-identity information notice, consult the US INFORM annual verification guide separately, distinguishing account-information tasks from product-content requirements. Changing IP addresses cannot alter purchasing eligibility, asset authenticity or platform permissions.
Sources
Frequently Asked Questions
Does TikTok Shop US prohibit all AI product images?
No. Official rules allow AI images that accurately represent the physical product. The sold item, PDP, SKU and image must match.
Can I use exaggerated effects if I add an AI label?
No. Production-method disclosure and product accuracy are separate checks. Misleading changes to dimensions, accessories, quantities, materials or functions still carry risk.
Can every PDP image use the video AI toggle?
That cannot be inferred. The general rules describe content-disclosure methods. Check the actual current entry point for product-detail images instead of inventing a universal toggle.
Do the September 24 rules apply to Malaysia and Vietnam?
The document verified here explicitly applies to the United States. Read each other market's local rules separately; do not infer same-day global enforcement.
What should I do if an image conflicts with a SKU?
Stop uploading, retain the original asset and evidence, rework it or use compliant physical-product photos, and recheck every SKU. Rolling back an asset is not guaranteed to cancel an existing review or penalty.