Check Fraud Trends in 2026: Why Financial Institutions Need AI
Despite years of predictions about paper checks being on the way out, check fraud now accounts for 30% of all fraud losses in the US (second only to debit card fraud). Between 2021 and 2023, suspicious activity reports for check fraud increased by 90%.
The latest headlines tell us just how common check fraud is. The Memphis Police Department recently charged a woman for depositing a stolen plumbing company check altered and cashed for over $6M. In another case, the Secret Service helped disrupt a NYC criminal syndicate that used bogus checks to steal over $20M in construction materials. And in what might be the boldest attempt, four men in Florida tried to cash a U.S. Treasury check worth $27 million.
What’s clear is check fraud isn't slowing down. And banks that don't adapt their detection strategies will continue losing millions while fraudsters accelerate ahead.
In this article, we're looking at the trends that surged check fraud in 2025 and how they're likely to intensify in 2026. More importantly, we'll explore how AI can help financial institutions strengthen their defenses against these methodologies.
Trend #1: Mail Theft Remains the Easiest Entry Point
Mail theft continues to be the preferred entry point for check fraud, and for good reason: it works. Organized criminals target USPS collection boxes, residential mailboxes, and increasingly, postal carriers themselves. The theft of USPS "arrow keys" has become particularly problematic. These keys provide access to multiple mailboxes and collection boxes across entire neighborhoods.
The insider threat is real and growing. In Torrance, a postal worker received a five-year prison sentence for stealing checks and cards from the mail over three years. A Florida mail carrier was arrested trying to sell USPS arrow keys and nearly $550,000 in stolen checks to undercover agents. And federal prosecutors charged another postal employee with stealing more than $1.6 million in checks.
Once stolen, checks enter a sophisticated criminal ecosystem. They're photographed, altered, and distributed. The mail system, designed for an era when physical security was the primary concern, remains vulnerable in ways that digital payments aren't.
The numbers confirm what law enforcement has been seeing on the ground. Between February and August 2023 alone, financial institutions reported $688 million in suspicious activity tied to mail theft-related check fraud. During that same period, FinCEN received 15,417 Bank Secrecy Act reports citing mail theft-related check fraud.
The United States Postal Inspection Service recovers over $1 billion in fraudulent checks and money orders annually. They've documented more than 38,000 high-volume mail theft incidents from mail receptacles, including those distinctive blue USPS collection boxes that sit on street corners across America.
How AI Can Prevent Mail Theft Check Fraud
While mail theft itself can’t be fully prevented, banks must deploy best-in-class fraud detection capabilities to stop stolen checks from turning into financial losses.
Anomaly Detection Models: These models work by establishing a baseline of normal behavior for each account: typical transaction locations, usual transaction times, standard deposit amounts, and regular spending patterns. When activity deviates significantly from this baseline, the system flags it for review.
Typology-specific models: These models are designed to detect known check fraud patterns, taking information such as how new a customer's bank account is, large check deposit and then attempting to send the value of that check out of the account.
Cross-Rail and Multi-Channel Insights: Fraudsters often exploit multiple payment rails (ACH, wire, checks, cards) simultaneously when committing fraud. AI models that integrate data across payment types and channels can detect coordinated attacks that might appear innocent when looked at in isolation.
Trend #2: Expansion of Black-Market Marketplaces
The business model of check fraud has fundamentally changed. Stolen checks are no longer used once by the thief who stole them. Instead, they're photographed and distributed through organized marketplaces on dark web platforms and encrypted messaging apps. It's the commercialization of check fraud. Just as legitimate businesses use marketplaces to reach more customers, criminals use Telegram channels and dark web forums to reach more buyers. And they've gotten efficient at it.
The speed is what makes this particularly challenging for banks. Half of stolen check images are posted on dark web platforms within eight days of theft, according to research from Recorded Future. By the time a victim realizes their check never arrived, it may have already been sold, altered, and deposited multiple times.
The scale is staggering. Researchers tracked 1.9 million stolen U.S. bank checks posted across over 700 Telegram channels in 2024. Nearly 1 million stolen bank check images appeared on dark and clear web platforms that same year.
How AI Can Help Prevent Black Market Check Fraud
Computer Vision Models: These models analyze signatures by calculating their mathematical "shape" and stroke vectors. By comparing this geometry against a known profile, the AI detects subtle deviations in size and style.
YOLO (You Only Look Once): This deep Learning model is a "single-shot" detector, meaning it looks at the entire image once to instantly predict bounding box coordinates and class probabilities. It handles the "where is the signature?" part of the task.
Entity graphs: This structured map of relationships between accounts, devices, and deposit locations allows banks to identify fraud rings operating across multiple accounts or institutions, tracing connections between seemingly unrelated fraudulent transactions that human investigators might not connect.
Trend #3: Mobile Remote Deposit Capture (MRDC) Exploitation
Mobile Remote Deposit Capture (MRDC) was designed to make banking more convenient. Customers could deposit checks from anywhere by taking a photo with their phone without a trip to the branch.
But convenience for customers became opportunity for fraudsters.
MRDC enables fraudsters to deposit altered or stolen checks without ever facing a bank teller. There's no in-person interaction where subtle tells might raise suspicion, no human employee who might notice something off about the check or the depositor's behavior. This has made MRDC the preferred method for "double presentment" schemes, where the same check is deposited at multiple institutions, and has made it far easier to exploit checks that have been washed or digitally altered.
The problem is complicated further by how MRDC systems work. Counterfeit checks made with incorrect check stock can bypass detection in MRDC systems that sophisticated image analysis. The security features that might be obvious to a trained teller viewing a physical check can be impossible to verify from a mobile phone photo.
FinCEN analysis found that many fraudsters specifically avoided in-person transactions by using mobile deposit apps. The financial impact is substantial. In 2024, 65% of financial institutions reported check fraud through RDC, and 80% of organizations faced attempted fraud via mobile deposits.
How AI Can Help Prevent MRDC Exploitation
Behavioral Biometrics: Uses AI models to analyze human patterns such as typing rhythm, mouse precision, and even device tilt. By calculating the "shape" of how a user interacts with a screen, the AI distinguishes a legitimate human from automated fraud software. For example, a device lying perfectly flat while typing at a flawless, rhythmic speed flags the unusual behavior typical of a fraud ring.
Device Intelligence: Complements behavioral biometrics by tracking Device IDs, IP addresses, and geolocation. This allows analysts to uncover schemes operating from the same location or flag when legitimate credentials suddenly appear on an unfamiliar, "cold" device.
Trend #4: AI-Powered Forgeries and "Check Cooking"
While you already know the term check washing, have you heard of check cooking? Unlike traditional check washing, which requires physical access to a check and chemicals, digital check cooking can be done entirely on a computer. A single stolen check image can be manipulated to produce multiple fraudulent checks.
Fraudsters now take advantage of AI-generated fake IDs, advanced photo-editing tools, and digital manipulation techniques to create forgeries convincing enough to bypass traditional verification processes.
The technology criminals use isn’t exotic or expensive. With generative AI and a home printer at their disposal, fraudsters can produce highly sophisticated forgeries. It’s advanced to the point where technical barriers to entry have largely disappeared.
This democratization of fraud tools means banks now face threats from an entire spectrum, from individuals facing economic pressure who can commit check fraud from their kitchen table to large-scale criminal operations running coordinated schemes.
How AI Can Help Prevent Check Forgeries
Image forensics: Machine learning models trained on millions of check images can detect subtle signs of digital alteration that are invisible to human reviewers, and that basic automated systems might miss. Handwriting and printed text are compared against expected norms for payees, amounts, and signatures.
The placement, size, and alignment of key check elements are also verified for consistency, and security features are assessed for authenticity. These models can identify anomalies or mismatches, such as differences between numeric and written amounts, unusual fonts, or irregular layouts.
Stop Check Fraud with Hawk
Check fraud isn't slowing down. It's becoming more sophisticated, more organized, and more expensive for financial institutions that rely on outdated detection methods.
Legacy detection systems weren't built for today's threats. They can't analyze check images at scale, can't spot behavioral anomalies in real-time, and can't connect the dots across payment channels. That's exactly what modern fraudsters are counting on.
Hawk's fraud platform gives you what traditional systems can't: AI-powered image forensics that catch alterations invisible to the human eye, anomaly detection that flags suspicious behavior before fraud completes, and unified monitoring across all payment rails so nothing slips through the gaps.
Your customers expect their deposits to be safe. Your institution expects losses to be contained. Hawk delivers on both.
To find out how Hawk can strengthen your check fraud detection, visit here.