AI-Generated Damage Claims Trigger Retail Crackdown on Return Fraud

Fraud Techniques Grow More Accessible

Scott Tannen was reviewing a customer service ticket when something stopped him. A shopper had submitted photos claiming a set of Boll & Branch sheets arrived torn, but the rip did not look like anything cotton does when it frays. One of the images carried an AI watermark.

Tannen, CEO of the New Jersey-based bedding brand, pulled more recent tickets. Several others contained damage photos that appeared to be machine-generated. The products, as far as he could tell, had arrived intact. The incident, which Tannen later described publicly on LinkedIn, put a face on a problem that is quietly spreading through retail operations.

Unlike traditional return abuse, which often involved “wardrobing” or reusing receipts, the new tactic relies on fabricated visual proof. Consumers can generate convincing images of scratched electronics, cracked screens or torn fabric using free image-generation tools. Those images are then uploaded through standard return portals designed to streamline legitimate complaints.

The scale of the problem is significant. U.S. consumers returned nearly $1 trillion in merchandise in 2024, more than double the total from four years prior, forcing retailers to spend an estimated $200 billion annually to recover value from returned goods, according to McKinsey.

Against that backdrop, fraudulent claims are adding a distinct layer of financial pressure. Research from Riskified, based on an analysis of more than 1 million refund claims, found that refunds account for 1% to 2% of total sales dollars, with nearly 1 in 4 refund dollars linked.

Retailers are also confronting the unintended consequences of their own growth strategies. Free returns, instant refunds and minimal verification helped accelerate eCommerce adoption. Those same features now limit how aggressively companies can tighten controls without alienating loyal customers. Filtering out abusive behavior while preserving convenience has become a central operational challenge.

Retailers and Logistics Providers Deploy Countermeasures

In response, merchants are investing in AI-based detection systems that analyze claim patterns, customer history and image metadata before refunds are approved. Rather than relying solely on human agents to spot anomalies, companies are embedding machine learning models directly into returns workflows.

Reuters reported that a UPS subsidiary deployed AI-based inspection technology to identify counterfeit and fraudulent returns during the 2025 holiday season, a period when return volumes peak and manual review becomes impractical at scale. The deployment illustrates how logistics providers, not just retailers, are being drawn into the fraud-detection infrastructure.

Riskified’s Dynamic Returns product is designed to apply differentiated treatment based on a customer’s assessed risk profile, offering faster resolutions to trusted shoppers while flagging high-risk claims for additional review. The approach reflects a broader shift in the industry toward risk-based return policies rather than uniform rules applied across a retailer’s entire customer base.

The fraud problem is accelerating a wider rethinking of reverse logistics as a business function. McKinsey estimates the reverse-logistics services market to be worth up to $14 billion, representing a major opportunity for carriers and third-party logistics providers to integrate more deeply with retailers. That integration increasingly includes fraud-detection capabilities alongside traditional routing and processing services.

Narvar, a post-purchase platform, has outlined the commercial case for converting returns infrastructure from a cost center into a revenue-generating function by using data from return interactions to personalize future customer engagement. The approach positions fraud mitigation not as a defensive measure but as part of a broader data strategy.