The Credit Card Industry Braces for Fraud Impacts As AI Moves to Take Over Shopping

Agentic artificial intelligence is bringing the future of commerce to consumers’ doors, but it also may pave the way for bad actors to use the technology to commit more fraud.

Credit card issuers and merchants may also have more chargebacks to deal with when AI agents begin managing a higher number of shopping tasks. But any increases in chargebacks may die down after a few years.

A new report from American Banker highlights that chargebacks will slow down over time after initially rapidly rising with the onset of agentic commerce.

Overall, global chargeback volume may grow to reach 324 million transactions in 2028 after coming in at 261 million in 2025, according to a prediction in a recent study from Mastercard and Datos Insights.

“It’s going to be messy for the next five years,” Armen Najarian, Chief Marketing Officer at Sift, a fraud-decisioning company that works with AI, told American Banker.

And agentic AI may be a catalyst to drive increases in fraud over the back half of the decade.

A recent report estimates that chargebacks will increase significantly from 2025 to 2028.

Speaking on a panel at the recent Money20/20 conference in Las Vegas, Nash Ali, Head of Operational Strategy at Experian, said that programs already exist that bad actors can use to leverage agentic AI to carry out fraudulent transactions.

“We’re sitting on the precipice now of another explosion in fraud with agentic AI coming our way,” Ali said.

The pace of fraud may be difficult for stakeholders in the credit card industry to keep up with when it’s amplified by the power of artificial intelligence.

“It’s no longer a human sitting and committing fraud on an individual basis,” Ali added.

Predictions for escalations in fraud may make credit card companies and merchants more cautious in the coming years when it comes to dealing with AI-powered shopping bots. But the payments industry may one day have agentic commerce to thank for decreasing levels of fraud.

Leaders in the credit card space, including Visa, have introduced solutions aimed at enabling secure communications between merchants and AI agents. 

Merchants may fear that they will have to pour significant resources into updating their systems to allow them to safely participate in agentic commerce. 

But Visa’s Trusted Agent Protocol, which the company rolled out last month, provides a way for merchants to verify the credentials of trusted agents without having to completely revamp their systems, according to a release from Visa on the solution.

Mastercard and PayPal have also recently released programs to support agentic commerce.

And Najarian indicated to American Banker that agentic commerce has the potential to greatly reduce incidents of friendly fraud, which occurs when a customer disputes a legitimate transaction as opposed to going to the merchant for a refund.

Shoppers are hesitant about permitting an AI agent to shop on their behalf.

Some consumers are reluctant to allow AI agents to make purchases for them, which may buy time for merchants and credit card companies to fine-tune their agentic commerce strategies on a smaller group of shoppers.

A new study from Sift shows that only 35% of consumers trust an AI agent to manage their financial information securely and make purchases for them.

But more people will likely begin using an AI agent for shopping as they gain confidence in the security and benefits of the technology.

Merchants, credit card issuers, and payment networks that work with one another to fend off fraud stemming from agentic commerce may be in a better position to weather initial challenges as the use of AI-powered shopping bots expands.