Fraud that targets payments is growing more complex and sophisticated. The criminal use of AI and the increasing standardization and reproduction of malware and other technologies: They act like a digital equivalent of old-fashioned burglary tool.
That’s according to the latest report released today by the threat intelligence experts at Recorded Future, which illustrates the cat-and-mouse game between bad actors and the banks, payment networks, governments and merchants that are trying to foil them.
On one hand, Recorded Future found that exposure of the vital information that payment fraudsters use in their scams flatlined last year. For example, the number of stolen credit card records accessible for sale dropped by almost 20% in 2025 from the year before. However, the increasing sophistication and scale of attacks mean that nefarious players can make ever better use of what they can access.
”In 2026, the organizations that succeed in defending their customers will be those that align leadership, fuse cyber and fraud intelligence, and maximize signal detection and correlation across fragmented data sources,” says Ilya Volovik, the company’s director of Payment Fraud Intelligence. “That’s how they can disrupt attacks before financial losses occur.”
One alarming trend that Recorded Future called out is the increasing industrialization of support services and technology that allows fraudsters to maximize their effectiveness. One such popular tool allows for scam operators to rewrite the code on an online retailer’s payments page and then steal, or “skim,” payment information as transactions happen in attacks called Magecart. The report points out that there were 10,500 such hacks active in 2025, leading to the compromise of over 23 million online transactions.
Tools are also increasingly available to help thieves steal the one-time passwords that banks and retailers use to validate identity and that typically bypass stronger forms of authentication that can foil fraud. Scammers are even figuring out how to automate the process of setting up their own fake retailers, allowing them to trick consumers into making a purchase they think is legitimate and thus handing over payment credentials.
Even old-fashioned paper checks aren’t immune: Thieves steal checks and post images online, exposing bank account information and signatures, and while the total number of images of stolen checks decreased by 42% compared to 2024, the number of unique images rose for the first time, to 233,000.
AI plays an increasing role in the scale and scope of payments fraud, according to the report. In one observed fraud, scammers used an AI-powered marketing platform to better target victims the way a legitimate business might try to lock in their own customer segmentation.
At the same time, the public is starting to use AI in ways that expand the opportunities for fraudsters. Although agentic commerce, in which an AI agent shops on behalf of a human user, is still in its pilot phase for many banks and merchants, the report says it could introduce new vulnerabilities, complicate investigations and raise liability issues. Knowing what’s at stake, the industry has already developed standards, protocols and tools to ensure trusted, transparent and secure agentic transactions.
The news isn’t all bad, though, and Recorded Future thinks the payments industry’s defenses will improve this year. As scammers deploy AI in more ways, predictive defenses will take precedence over reactive ones. The growing standardization of fraud tools, while helping scammers, also, ironically, presents an opportunity for the good guys. Those payment information “skimming” kits that allow for infiltration of retailers’ websites are now so widespread that indicators of scams now surface other similar instances and could soon simplify detection.
Financial institutions and payments organizations have many tools of their own with which to combat fraud but perhaps the most effective is to embrace a posture that favors heading off threats before they grow into large-scale attacks. Coordinating between intelligence and fraud detection, using technology to speed response time and involving organizations' leaders will lead to better, faster decisions and the ability to disrupt fraud before it leads to financial losses.
