Detect Deepfakesby Resemble AI
Deepfake case study · Image

AI Makes Refund Evidence Easier to Fake - Practical…

Retailers like Bogg Bag and Boll & Branch face rising ecommerce refund fraud as generative AI tools enable criminals to create convincing fake evidence

Incident date
Jul 2026
Target
Bogg Bag, Boll & Branch
Updated Jul 17, 2026 · 1 min read

Ecommerce retailers are increasingly vulnerable to sophisticated refund fraud as generative AI tools allow bad actors to produce highly realistic, forged evidence. Retailers such as Bogg Bag and Boll & Branch have already reported encounters with AI-falsified refund proof, highlighting a growing trend that threatens to undermine the trust-based systems used by online merchants to process returns.

What happened

Online merchants typically approve refund requests based on provided documentation, such as customer descriptions and photographs of damaged goods, without physically inspecting the items. This process, often utilized for inexpensive or perishable products to save on shipping and handling costs, relies on the assumption that the submitted evidence is authentic. Generative AI has rendered this assumption obsolete. Criminals are now using simple text prompts to create plausible imagery of product defects, including cracks, stains, mold, tears, and dents, as well as damaged packaging.

Beyond image manipulation, AI is being leveraged to manufacture a complete deceptive narrative. Fraudsters use these tools to fabricate shipping records, carrier documents, delivery screenshots, and customer service correspondence that implies a merchant has already approved a refund. Because these tools can generate multiple iterations of a claim with minimal effort, the fraud is highly scalable and can be automated across numerous accounts and merchants. While retailers are exploring defensive measures—such as reviewing image metadata, conducting reverse-image searches, and implementing manual reviews for high-value claims—these interventions often introduce significant operational costs and potential customer friction. As detection tools struggle to keep pace with the rapid advancement of image generation, businesses are finding that the ease of creating synthetic evidence poses a significant challenge to traditional loss prevention.

Sources