Artificial intelligence shopping tools used by Amazon and Walmart can reportedly detect misleading Made in USA claims but often fail to warn shoppers about them, according to a new study.
Researchers tested Amazon’s Alexa for Shopping and Walmart’s Sparky by asking questions about products that advertised American manufacturing in their titles while listing imported origins elsewhere in their specifications.
The study found that the systems could understand the contradiction. In several cases, the product title claimed it was made in the United States, while less visible listing details showed that it came from another country.
Despite recognising the conflict, the AI assistants did not consistently bring it to the buyer’s attention.
This distinction is important because it suggests the problem may not come from weak data processing. The systems appear capable of reading the relevant information, but their responses do not always reveal it clearly.
The study found different treatment for origin questions
The researchers reported that Amazon’s shopping assistant sometimes blocked or limited questions about products claiming to be made in the United States.
Similar questions about goods made in China reportedly received more detailed answers.
When asked about the difference, the AI assistant suggested that the restriction was connected to company policy and the potential effect on overseas sellers. However, responses generated by a chatbot should not automatically be treated as an official explanation of corporate policy.
The study argues that the wider pattern still shows selective transparency. Both retail platforms hold information about product origins, and their AI tools can process it, yet shoppers may not receive a clear warning when listings contain conflicting claims.
| Issue examined | Reported finding |
|---|---|
| Product titles | Some claimed the item was made in the United States |
| Specification details | Some identified an imported country of origin |
| AI understanding | The assistants could detect the contradiction |
| Shopper warning | The conflict was not consistently disclosed |
| Amazon responses | Some Made in USA questions were limited or blocked |
| Walmart responses | Sparky also failed to reliably flag conflicting listings |
| Main concern | Buyers may receive incomplete information before purchasing |
False origin claims can affect buyers and domestic businesses
Some shoppers are willing to pay more for products manufactured in the United States. Their reasons may include supporting domestic employment, avoiding certain supply chains, or meeting purchasing requirements.
A misleading label can therefore influence both the purchasing decision and the price a customer accepts.
False claims can also disadvantage companies that genuinely manufacture their products in the United States. Domestic production may involve higher labour, material, and regulatory costs. A seller that imports a product but markets it as American made can compete unfairly against businesses following the rules.
Federal guidance generally requires an unqualified Made in USA claim to mean that all or nearly all of the product’s significant parts and manufacturing work come from the United States.
Products that contain substantial imported content may need more limited wording that accurately explains where they were assembled or which parts came from the country.
Marketplace listings often contain conflicting information
Large online marketplaces receive product data from many sellers. Titles, descriptions, technical specifications, images, and backend fields may be entered separately.
This can create inconsistencies, whether through mistakes or deliberate attempts to improve visibility in search results.
AI shopping assistants could help identify these conflicts before a buyer completes a purchase. A useful warning might state that the title claims American production while the specification field lists another country.
Such a system would not need to decide whether fraud occurred. It could simply present the conflicting information and allow the shopper to make an informed decision.
The study suggests that the platforms are not using their AI tools consistently for that purpose.
Researchers want greater accountability around retail AI
The report also raises concerns about how companies document the behaviour of consumer facing AI systems.
Amazon and Walmart have previously discussed responsible AI development and transparency, but the researchers argue that shoppers still lack enough information about how product answers are filtered or restricted.
They call for stronger oversight and clearer disclosure of the rules governing shopping assistants. The report also suggests that regulators and other authorities may need to examine whether misleading origin claims are being handled properly.
The study does not establish that every Made in USA listing on either platform is false. It focuses on examples where visible marketing language conflicted with other product information.
Its central concern is that AI tools may know more than they tell shoppers.
As conversational assistants become a larger part of online retail, buyers may rely on their summaries instead of reading every specification field. That makes accuracy and full disclosure increasingly important. An assistant that overlooks conflicting origin information can make a misleading listing appear more trustworthy than it is.
Discussion (0)
Be the first to comment.