The news: AI is simultaneously making returns fraud easier to commit and easier to prevent.

• More retailers are using AI tools from companies like Happy Returns, Narvar, and Loop to detect fraudulent returns and drive other improvements in their reverse logistics capabilities, per Business Insider.
• But more consumers are using AI to generate their return and refund requests, as well as create fake images, receipts, and other evidence to bolster fraudulent claims.

The big picture: AI has the potential to help retailers solve major pain points related to returns. Narvar claims that its tools can help identify and block fraudsters, reduce claim payouts, and protect retailers’ revenues and customer relationships. Others, like Loop, use data from millions of customers and purchases to flag suspicious behavior and identify “high risk” shoppers who are likely to be frequent returners.

The potential cost savings are substantial: Roughly 9% of returns are fraudulent, according to a 2025 report by the National Retail Federation and Happy Returns. Based on our forecast, that could amount to roughly $78.9 million in lost revenues this year alone. It’s no surprise, then, that implementing AI to predict and prevent returns ranks among worldwide ecommerce retailers’ top reverse logistics initiatives, with nearly 2 in 5 (39%) expecting to do so this year, according to Asendia.

However, companies aren’t the only ones getting more sophisticated. Roughly half of consumers now use AI tools like ChatGPT or Claude to help prepare return or refund claims, according to a Riskified report. That trend, coupled with shoppers’ tendencies to stretch the limits of retailers’ returns policies, could eventually make it more difficult for brands to differentiate legitimate claims from fraudulent ones.

Brands like Boll & Branch and Bogg are already encountering fraudsters who are using AI to alter images to show product damage in hopes of securing a refund. Other tactics include faking shipping receipts that falsely indicate returns were dropped off at UPS or USPS, or fabricating police reports to claim items were stolen, according to Modern Retail.

Implications for retail: AI-driven returns fraud poses broader challenges to retailers’ bottom lines as the technology goes mainstream. In particular, the rise of agentic commerce could erode the effectiveness of retailers’ existing fraud detection systems: A recent report from Chargeback911 found that payment providers’ current tools are inadequate at detecting AI-enabled chargeback fraud, especially when autonomous agents are involved.

With more traffic coming from bots, retailers have fewer ways to distinguish legitimate shoppers from bad actors—a problem that will only grow as more consumers tap agents to buy on their behalf.