The rise of artificial intelligence has fundamentally altered the landscape of workplace ethics, turning the once-arduous task of forging an expense receipt into a trivial, seconds-long endeavor. According to a recent survey by expense management platform Emburse, a staggering four in ten U.S. employees have utilized AI to manipulate their expense reports. This isn’t limited to just “doctoring” numbers; 19% of respondents admitted to fabricating expenses for purchases that never occurred, while others inflated existing receipts or used AI to recreate lost documentation. Perhaps most startling is the ease with which this is happening: 40% of these individuals are actually using company-funded AI tools to carry out these deceptions, signaling a massive blind spot in current corporate governance.
The data reveals that this isn’t merely a fringe problem—it is a trend that is exploding in scale. According to figures from AppZen, the number of AI-generated receipts flagged as fraudulent across their platform surged from 0% in early 2025 to over 70% by mid-2026. This surge represents thousands of fraudulent entries across hundreds of companies, totaling hundreds of thousands of dollars in illicit reimbursements. Unlike the clumsy, template-based forgeries of the past, these AI creations are designed to fly under the radar. By keeping the fraudulent amounts relatively low—averaging around $100—employees are skillfully maneuvering below the auto-approval thresholds that many corporations rely on to streamline their finance workflows, effectively rendering traditional manual reviews largely ineffective.
The core of this issue lies in the fact that our traditional systems for corporate accountability are now obsolete. For decades, the receipt served as a trusted proxy for the truth; if it looked authentic, it was assumed to be real. However, AI generators can now produce high-fidelity images complete with convincing paper textures, precise itemization, and accurate timestamps in under a minute. As Chris Juneau from SAP Concur pointed out, the quality of these fakes is now so sophisticated that finance teams have been warned: “Do not trust your eyes.” Because these documents are generated from scratch rather than edited, they lack the tell-tale visual “artifacts” of older forgeries, such as pixelation or inconsistent fonts, making them indistinguishable from legitimate records through visual inspection alone.
Crucially, this wave of dishonesty is often fueled by personal financial desperation rather than simple malice. Emburse’s findings indicate that over half of U.S. employees have suffered from overdraft fees, late-payment charges, or credit card interest—often because they are left waiting for their own companies to process legitimate reimbursements. When employees feel consistently behind on their own bills, the temptation to categorize a personal purchase as a business expense or to fabricate a receipt to cover a shortfall becomes a survival mechanism. As Michele Shepard from Emburse notes, most employees aren’t trying to be “bad actors”; they are simply responding to a high-pressure financial environment using the new tools they’ve been encouraged to adopt, leaving employers struggling with a significant lack of visibility.
Because AI-generated documents look perfect to the human eye, the only defense left is to move away from document-based verification entirely. Attempting to catch a fake by examining the receipt is now a losing battle; the focus must shift to verifying the transaction itself. Companies must move toward data-matching protocols that reconcile submitted receipts against actual bank card transaction data, merchant category codes, and real-time payment timestamps. If the numbers don’t match the bank’s digital audit trail, the receipt is effectively useless, regardless of how professional it looks. Verification is no longer about checking paper textures; it is about cross-referencing activity in real time.
Ultimately, the solution to this crisis requires a blend of technology and culture. As industry leaders like Ernest Rolfson suggest, the only way forward is to build a proactive defense that validates identities and payment instructions at every stage of the transaction process. Companies are being urged to adopt virtual cards, which link specific payments to singular, pre-authorized transactions, making it physically impossible to submit an unsupported receipt. By combining these rigorous digital controls with clear, transparent expense policies and—most importantly—speedier reimbursement times for honest employees, organizations can bridge the trust gap. In the age of AI, the receipt is no longer proof of truth; only the transaction data itself can hold that weight.