Payment Providers Face Liability Question as AI Agents Gain Transaction Power

Financial institutions are bracing for a surge in fraud as artificial intelligence systems gain the ability to handle transactions autonomously. Roughly 78% of firms anticipate a significant uptick in fraudulent activity tied to agentic commerce, according to data from xpate, a global payments and banking platform.

The preparedness gap is equally stark: 85% of institutions acknowledge their current infrastructure and growth strategies fall short of what's needed to manage risks in this emerging landscape. The figures were released during Cybersecurity Awareness Month.

Sofia Khatsernova, legal function owner at xpate, maintains that granting decision-making authority to AI in financial contexts demands embedded safeguards from the outset. Without predetermined limits, the technology poses unacceptable legal and operational hazards.

You can't give AI autonomy in finance without building in constraints. Legal and compliance teams will need to define clear boundaries around what these systems can do, how decisions are documented, and where liability sits, Khatsernova said.

Khatsernova argues that traditional consent frameworks — a single checkbox at account creation — are inadequate when software makes ongoing decisions for customers. The paradigm must shift toward continuous, granular authorisation.

Consent also needs to evolve, not as a one-time agreement, but as something more dynamic and granular, with built-in safeguards that prevent autonomous systems from acting beyond the authority users have granted them, she said.

This approach requires payment platforms to redesign how user permissions are captured, updated, and enforced throughout the customer relationship.

Payment providers confront a trio of practical challenges as agent-initiated transactions become routine:

• Establishing explicit limits on the types and values of payments an AI agent can trigger
• Creating audit trails that capture the logic and data behind each autonomous decision
• Determining liability allocation when an agent executes an unauthorised or erroneous payment

Firms that resolve these questions early — before regulatory frameworks solidify — stand to gain competitive advantage by building trust with both oversight bodies and end users.

As AI-powered payment systems transition from experimental pilots to mainstream deployment, trust becomes the critical currency. Customers and regulators alike will scrutinise how providers handle accountability when machines act independently.

The institutions that proactively address consent granularity, decision transparency, and liability frameworks are positioning themselves to navigate the regulatory landscape — and customer expectations — more smoothly than those treating agent autonomy as a purely technical challenge.