Story perspectives

Banks Warn AI Shopping Agents Could Boost Fraud Risks

By Drooid · Sep 23, 2026, 10:20 PM UTC · How we work

Summary

Banks Warn AI Agents

Six banks including NatWest and Bank of America warned AI shopping agents could increase scams, fraud and data breaches.
The banks said agents may request card details and enter them on merchant sites.
They urged mandatory AI-agent transaction disclosures and transparent decision-making.
John Lewis reported AI-originated searches rose to 2.5% from 0.3% a year earlier.
Banks will discuss proposals with policymakers to establish trust and data safeguards.

Summary

Banks Fund AI

Banks have lent roughly $800 billion to AI data-center and infrastructure projects, per Columbia professor Stijn Van Nieuwerburgh.
If AI firms default, banks could lose capital and restrict lending to the broader economy.
Many community banks rely on a single core-software vendor, exposing them to coordinated rogue AI attacks.
Effective AI governance requires deployment inventories, clear ownership, mandatory human review of high-risk decisions, and observability of agents’ chain-of-thought reports.

History

Historical Context

Lehman Brothers’ September 2008 collapse sparked a $700 billion bailout and a wave of reforms. In June 2009 President Barack Obama’s team proposed new rules, and Congress signed the Dodd-Frank Act in July 2010, creating the Consumer Financial Protection Bureau. Regulators now cite that history while weighing AI-driven lending disclosures to guard credit exposure.

1 of 3

Banks Warn AI Agents

Six banks including NatWest and Bank of America warned AI shopping agents could increase scams, fraud and data breaches.
The banks said agents may request card details and enter them on merchant sites.
They urged mandatory AI-agent transaction disclosures and transparent decision-making.
John Lewis reported AI-originated searches rose to 2.5% from 0.3% a year earlier.
Banks will discuss proposals with policymakers to establish trust and data safeguards.

1 / 3