Instant payments shift fraud liability to real-time
Beyond the victims and billions of dollars lost, the rise of AI-driven fraud means banks, fintechs and payments providers must consider risk, liability and real-time money flows.
Fraudsters are using the same intelligence tools used in modern business to increase the speed, scale and success of their attacks. The result is a growing gap between the speed at which fraud develops and the slowness with which traditional chargeback and liability frameworks respond.
Refund and return rules are not designed to target AI fraud
This faster, more adaptable threat landscape exposes a growing mismatch between traditional chargeback eligibility rules and the reality of AI-driven fraud. Banks have historically adjudicated chargebacks based on static criteria related to transaction type, customer history, or manual review.
But now, as fraud moves toward rapid credential compromise and unauthorized use, these frameworks are coming under pressure.
As detailed in the State of Fraud and Financial Crime in the U.S. 2025 report commissioned by Block and released by PYMNTS Intelligence, agencies report that fraud is no longer just a financial loss issue.
According to the report, 50% of financial institutions said fraud has eroded customer loyalty, 44% reported damage to their brand and reputation, and 48% reported lost business opportunities.
AD: SCROLL TO CONTINUE
These trust-based consequences increase pressure to make chargeback decisions quickly and consistently, especially when consumers expect immediate resolution in a context where funds can be transferred instantly between accounts.
AI-driven fraud accelerates across channels
Fraudsters are now weaponizing artificial intelligence to perpetrate fraud through deepfakes, impersonation and exploiting personal data. These attacks have directly contributed to a surge in unauthorized party fraud, which now accounts for 71% of all fraud incidents and dollar losses, reversing last year's pattern in which authorized party manipulation was more common.
Data from The Block and PYMNTS Intelligence shows that credential compromise, account takeovers and impersonation scams have risen sharply, with digital payment fraud accounting for 20.3% of total fraud losses on a dollar-weighted basis.
These patterns demonstrate how AI can enhance the ability to mimic legitimate identity signals or automate high-volume attacks that bypass static controls.
Instant money movement makes liability a real-time issue
The proliferation of instant payments has made these tensions even more acute. Real-time money movements mean that institutions must “score” trades before they settle, rather than as late as in the past (sometimes hours).
Operational pressures and challenges are evident in the data. The report found that 46% of institutions said faster payments were the top fraud management challenge, and 41% pointed to the expansion of payment types and currencies, including peer-to-peer and instant transfers.
This places liability issues squarely in the transaction process, with institutions having to authenticate participants, verify operations and determine whether customers are eligible for compensation if the transaction is later deemed fraudulent.
Artificial intelligence and behavioral analytics provide a line of defense
To keep pace, financial institutions are rapidly modernizing their fraud defenses. The PYMNTS Intelligence/Block report shows that 70% of organizations currently use behavioral analytics, while 61% use machine learning.
These systems check for behavioral signals that would be difficult for fraudsters to deceive, even with artificial intelligence tools. Device fingerprinting, speed checks, transaction context and behavioral baselines help determine whether transactions match historical patterns of legitimate customers.
The report also shows how embedded artificial intelligence can transform decision-making. 70% of agencies say machine learning can help them balance both proactive and reactive strategies, and 25% say it makes their defenses more proactive.
This reflects a shift toward instant rail real-time detection, with the goal of intercepting fraud attempts before funds leave an account.
These initiatives demonstrate a recognition that manual and rules-based interventions cannot support high-speed chargebacks and liability determinations for instant transactions.
Expectations around fraud solutions are changing rapidly. Faster payments reduce the time it takes for agencies to detect, investigate and decide whether a client should be reimbursed. Artificial intelligence and behavioral analytics solve this problem by strengthening identity assurance and enabling early signals of unusual activity.
Institutions that modernize their defenses can not only reduce fraud losses but also increase customer confidence. The path forward is clear: Real-time payments require real-time fraud intelligence, and institutions that deeply integrate AI into risk scoring, identity proofing, and chargeback governance will be able to better control liability and maintain trust.
