Columbia Study Says Retail AI Detects but Does Not Flag Origin Conflicts

Columbia Law School's Center for Law and the Economy reported on July 30, 2026, that Amazon's Alexa for Shopping and Walmart's Sparky can detect contradictions in some 'Made in USA' listings but do not flag those discrepancies for shoppers. The study, reported by Reuters, followed the Federal Trade Commission's July 2025 call for both marketplaces to police potentially deceptive origin claims from third-party sellers. Walmart's Sparky responded that the FTC generally enforces Made in USA rules against manufacturers rather than retailers, calling it 'a business calculation, not a legal justification.'

Study Finds Retail AI Misses Origin Fraud Columbia Law School researchers reported on July 30, 2026, that Amazon's Alexa for Shopping and Walmart's Sparky can detect contradictions in some Made in USA listings but do not flag those listings. Reuters reported that the findings followed the FTC's 2025 call for both marketplaces to police potentially deceptive origin claims from third-party sellers. Amazon's Alexa for Shopping and Walmart's Sparky can identify conflicts between "Made in USA" claims and country-of-origin information within product listings, but neither assistant flags those discrepancies for shoppers, according to a report from Columbia Law School's Center for Law and the Economy. Reuters reported on July 30 that the study examined how the two retail chatbots handled product-origin queries and misleading labeling. Under the Federal Trade Commission's Made in USA Labeling Rule, products using the label must be "all or virtually all" made in the United States. What the researchers found The Columbia researchers reported that both assistants could recognize when an explicit U.S.-origin claim in a title or description conflicted with origin details elsewhere on the same listing. The report characterized such conflicts as potentially deceptive "Made in USA" claims. According to Reuters and Quartz, researchers also asked the chatbots why suspicious claims were not flagged. Walmart's Sparky responded that the FTC generally enforces Made in USA rules against manufacturers rather than retailers, adding: "That's a business calculation, not a legal justification." The study also reported that Amazon restricted some responses to queries for U.S.-made products while allowing comparable China-origin searches. Those chatbot responses are findings reported by the researchers, not independently verified statements of Amazon or Walmart policy. Retailer and regulatory responses An Amazon spokesperson told Reuters that country-of-origin information is displayed on product detail pages when available and that Amazon is continuing to improve Alexa for Shopping's ability to provide that information accurately. Amazon also told Reuters that it acts when it finds sellers violating its policies. Walmart did not immediately comment to Reuters. Retail Gazette likewise reported that Walmart had not immediately responded to the findings. The FTC wrote to Amazon and Walmart in July 2025 urging the companies to address potentially deceptive Made in USA claims by third-party sellers, according to Reuters and Retail Gazette. The agency's rule can expose companies that make unsupported origin claims to legal action and financial penalties. Detection is not enforcement The report illustrates a recurring distinction in AI-enabled marketplace governance: a system's ability to retrieve or reason over contradictory listing data is different from an operational decision to label, demote, remove, or investigate a listing. In comparable marketplace systems, converting detection into enforcement requires confidence thresholds, evidence retention, appeal processes, and policies that account for seller and consumer harm. For ML and trust-and-safety teams, the case also underscores that chatbot behavior should not be treated as a complete audit of platform controls. A conversational assistant may surface information from product metadata, while the marketplace's actual moderation pipeline can operate through separate rules, human review, and seller-enforcement processes. The Columbia study raises questions about whether those systems are being connected effectively for origin-labeling claims. Key Points - 1Columbia researchers found Alexa and Sparky could detect origin-data contradictions, yet the listings were not flagged. - 2FTC rules require products labeled "Made in USA" to be all or virtually all domestically made, raising marketplace compliance stakes. - 3Comparable marketplace systems show that detection alone does not create enforcement without review workflows, evidentiary controls, and remediation policies. Scoring Rationale The report concerns AI-assisted marketplace governance at two of the largest U.S. retailers and connects model behavior to consumer-protection enforcement. It is particularly relevant to teams building product search, conversational commerce, and trust-and-safety workflows, although the findings are limited to a specific labeling category and study methodology. Sources Public references used for this report. Practice with real Ad Tech data 90 SQL & Python problems · 15 industry datasets Active Search Campaigns by BudgetEasy /problems/sql/active-search-campaigns-by-budget High CPC Clicks & Poor Landing PagesMedium /problems/sql/high-cpc-clicks-poor-landing-page Campaign ROAS by Attribution ModelHard /problems/sql/campaign-roas-by-attribution-model 250 free problems · No credit card See all Ad Tech problems /problems/datasets/adtech