Regula Report: Only 64% of Banks Can Reliably Distinguish Humans from AI Agents in Identity Flows

Quick Summary

Regula's 2026 global report reveals a critical identity verification gap: only 64% of banks can reliably prove a human is behind remote checks. As AI agents automate workflows, 52% of organizations struggle to verify biometric liveness, necessitating a shift from simple pass-fail results to continuous identity assurance.

How Do AI Agents Impact Identity Verification?

AI agents are no longer passive tools; they are becoming active participants in digital banking workflows. Regula’s research indicates that 69% of organizations already encounter AI tools within their identity verification flows. This shift means that automated systems can now mimic human behavior, submitting evidence and navigating onboarding journeys with high precision. For financial institutions, the challenge is no longer just spotting a fake document, but determining if the entity interacting with the system is a real human user or an automated script. The report highlights that while 76% of firms have technical controls for human presence detection, only 48% actually trust them to perform reliably against sophisticated AI-driven attacks.

What Are the Primary Risks to Biometric Security?

The rise of synthetic identity fraud and deepfakes has rendered traditional face matching insufficient. According to the study, 52% of organizations cannot fully verify if biometric data was captured live, while 41% struggle to detect manipulated identity signals. Key metrics from the report include:

• 64% of banks claim reliability in human detection, performing better than the global average.
• 52% of organizations fail to guarantee biometric liveness.
• 41% of businesses cannot detect signal manipulation in real-time.

These figures suggest that injected camera feeds and replay attacks are successfully bypassing legacy systems, forcing a move toward multi-signal identity assurance.

How Can Banks Maintain Trust in an AI-Driven Environment?

To secure remote identity checks, businesses must move beyond one-time onboarding scans. The report advocates for a continuous trust model that validates the integrity and origin of every signal throughout the customer lifecycle. As AI agents become more integrated into payments and trading, identity verification must evolve to analyze the interaction itself, not just the credentials provided. This involves verifying biometric authenticity in real-time and using forensic-level checks to ensure that identity evidence hasn't been synthetically generated or altered by AI tools.

"Identity verification was built for a world where every interaction involved a human being. That assumption no longer holds. Today, organizations must verify not only identity signals but also the interaction itself. They need to understand whether biometric data was captured live, whether identity evidence is authentic, and ultimately whether a real person is actually behind the session. The future of identity verification is not a single check performed during onboarding. Trust can no longer be established once and assumed forever," said Henry Patishman, Executive Vice President of Identity Verification Solutions at Regula.

FF NEWS TAKE:

This report from Regula is a wake-up call for the industry. The fact that nearly half of organizations don't trust their own identity verification tools while AI agents proliferate is a massive red flag. As the Bank of England eyes regulatory reform, banks must realize that 'good enough' security is dead. The industry is moving toward a reality where biometric liveness is the only true currency of trust, and those failing to adapt to automated identity threats will be left wide open to systemic fraud.