3AM Marketer
AI

Why Banks Are Blocking Unauthorized AI Agents

Financial institutions are blocking AI agents that rely on credential sharing, transitioning to read-only API access and FAPI standards to mitigate security risks.

4 major banks
5 sources, checked 26 Oct 2026.

Major U.S. financial institutions, including JP Morgan, Bank of America, Citi, and Morgan Stanley, have detected third-party AI agents operating inside customer accounts. The inability to distinguish between human and bot traffic is forcing institutions to block this technology and revise their terms of service.12

The logo of JPMorgan Chase, one of the financial institutions affected by unauthorized AI agent access.
JPMorgan Chase (then Chase Manhattan), Public domain, via Wikimedia Commons

The primary security challenge for financial institutions is attribution. When a customer provides login credentials to an AI agent, the agent becomes indistinguishable from the customer, leaving the bank unable to tell whether activity originates from a human user or an automated system.13

Why screen scraping creates critical vulnerabilities

Screen scraping accesses user data and functionality by impersonating a user through password sharing. The OpenID Foundation characterizes this practice as brittle, inefficient, and insecure because relying on shared passwords creates direct security vulnerabilities.4

Technical solutions: FAPI and DPoP

To secure third-party integrations, the industry is turning to hardened authentication specifications. The Financial-grade API (FAPI) security profile provides a higher level of security than standard OAuth or OpenID Connect for high-risk API use cases. RFC 9449 introduces Demonstrating Proof-of-Possession (DPoP), an application-level mechanism that enables an OAuth client to prove possession of a private key. DPoP stops unauthorized parties from using leaked or stolen access tokens by binding tokens to a public key upon issuance and verifying the client holds the matching private key, preventing replay attacks.45

The official OpenID Foundation specification page for the Financial-grade API security profile.
Screenshot of openid.net, captured 2026-10-26

How banks mitigate agent access risks

Financial institutions are shifting from raw credential access to constrained integration models to limit operational exposure.13

Access MethodCredential MechanismAttribution CapabilityRisk Profile
Screen ScrapingRaw password sharingIndistinguishable from human userHigh security vulnerabilities
More on Screen Scraping
Risk Profile
High security vulnerabilities
Read-Only API AccessConstrained API tokensSeparates viewing from transactional executionInterim risk mitigation
More on Read-Only API Access
Risk Profile
Interim risk mitigation
FAPI with DPoPPublic-private key bindingSender-constrained client verificationHigh security standard
More on FAPI with DPoP
Risk Profile
High security standard

Most banks have authorized or are currently moving toward read-only access that allows external tools to view financial data without executing transactions. Because full write access via customer credentials presents elevated risks, organizations must abandon credential-sharing automations and transition integration roadmaps to read-only APIs and sender-constrained token standards.13