Dojo warns of AI fake receipts fuelling refund fraud

JOSEPH GABRIEL LAGONSIN

News Editor

Dojo has warned that fraudsters are using AI-generated receipts to secure refunds from retailers, with online searches for such tools rising sharply.

Data cited by the payments company showed global Google searches for “AI-generated receipts” rose by 2753% over the past year. It also referenced findings from Cifas showing that AI-manipulated documents now account for more than 20% of falsified evidence used in UK refund and chargeback disputes, where total fraud cases exceeded 444,000 last year.

The issue centres on fake proof-of-purchase documents created with generative AI tools. These can mimic real receipts with retailer logos, item lists, timestamps and transaction numbers, making them hard for frontline staff to distinguish from genuine records during refund checks.

According to the figures cited, about one in 10 UK returns is already estimated to be fraudulent. AI-generated imagery is adding pressure to return, refund and replacement processes, particularly where staff rely on uploaded documents rather than direct checks against payment records.

A Dojo spokesperson said AI receipt generators are making it easier for fraudsters to create convincing fake proof of purchase.

“With just a few clicks, they can produce receipts that mirror real store formats, complete with logos, item lists, timestamps and transaction IDs. As many return systems rely on documentation rather than transaction-level checks, these AI-generated receipts often go unnoticed, allowing fraudulent refunds to slip through even well-established defences.”

Risk areas

Several sectors appear more exposed than others. Dojo identified eCommerce and digital-only retail, fast-food and quick-service restaurants, consumer technology, ride-share and on-demand travel, food-delivery platforms, luxury goods, telecoms and financial services as industries where interest in receipt generators suggests higher fraud risk.

The assessment was based on Google search patterns for brand-specific receipt generator terms. Separate Experian data for 2025, cited by Dojo, indicated that 62% of digital-only retailers, 48% of retail banks and 44% of telecom providers reported AI-related fraud attempts.

Those sectors share several features that can make them targets for document fraud. High transaction volumes, automated customer service systems and digital templates that are relatively easy to imitate can create openings for false refund, return or reimbursement claims.

Checks advised

To limit exposure, Dojo urged businesses to verify submitted receipts against actual transaction records. Checks on timestamps, payment amounts and transaction IDs remain the most reliable way to identify fabricated documents.

It also recommended adding QR codes, barcodes or other unique markers to receipts so staff can trace a claim back to a live record in the merchant’s system. For higher-risk categories such as expensive electronics, digital-only purchases and goods commonly resold online, retailers should apply stricter refund rules and require further verification before approving a claim.

Another step is staff training. Frontline workers can be taught to spot repeated textures, irregular spacing, mismatched metadata, incorrect store formats and impossible timestamps, all of which may indicate that a receipt or supporting image has been artificially generated or altered.

Behavioural monitoring is also part of the advice. Repeated refund requests from the same device or location, multiple claims in a short period, or unusual patterns in customer activity can help investigators identify serial abuse even when the documents appear plausible.

Dojo also advised businesses to consider AI-based fraud detection systems that examine images and documents for signs of manipulation, including synthetic text, altered pixels and inconsistencies that may not be visible to staff reviewing claims manually.

The warning reflects a broader shift in fraud methods as generative AI tools become easier to access. For merchants, the challenge is no longer limited to spotting obvious forgeries, but to checking whether documents submitted in routine customer service interactions match real transaction data held in their own systems.