AI agents are a digital identity headache despite explosive growth

AI agents are posing a challenge for digital identity assurance companies, both in terms of differentiating them from human identities and to ascribe their ownership and rightful limitations. A method of linking AI agents to biometrics has been developed by ZeroBiometrics, and Anetac has introduced a platform to increase visibility into identities it says can help.

MarketsandMarkets forecasts the global market for AI agents will grow from $5.25 billion in 2024 to $52.62 billion by 2030, a CAGR of 46.3 percent, in a report cited by Frontegg as it launched an identity layer. Organizations interacting with all of those AI agents will need to know whose they are and what they are authorized to do.

Okta Senior Emerging Tech Researcher Fei Liu told SC World as RSA Conference 2025 opened that “most organizations don’t have a strategy for managing” AI agents, despite this expected growth.

“Unlike traditional identities, AI agents need access to user-specific data and workflows to make decisions,” she notes.

AI agent identity was one of the key themes of the conference.

Template-free biometric binding with ZeroBiometrics

ZeroBiometrics proposes linking AI agents to the identities of real humans using its biometric cryptographic technology. This is possible due to a new system the company hails as groundbreaking, in which a short chain of certificates binds agents to humans and humans to their biometrics.

The proprietary, template-free biometric cryptography from ZeroBiometrics is used to issue root or intermediate certificates to humans. The individual can be verified in real time using privacy-preserving biometric authentication, after which they can issue end-entity certificates to their AI agents to allow them to carry out transactions within a set of pre-defined limits.

Certificates can attest to the identity of either an individual or an organization.

“AI agents are becoming more powerful, but without trust anchors, they can be hijacked or abused,” says Alfred Chan, CEO of ZeroBiometrics. “Our technology ensures that every AI action can be traced to a real, authenticated person—who approved it, scoped it, and can revoke it.”

ZeroBiometrics says its new AI agent solution makes use of open standards and technology, and supports transaction controls including time limits, financial caps, functional scopes and revocable keys. It can be integrated with decentralized ledgers or PKI infrastructures, and is suggested for applications in finance, healthcare, logistics and government services.

Identity layer launched by Frontegg

The lack of identity standards suited to AI agents is creating a major roadblock for developers trying to address the looming market, according to Frontegg.

That is why it has developed an identity management platform for developers building AI agents, saving them from spending time building ad-hoc authentication workflows, security frameworks and integration mechanisms. Frontegg’s own developers discovered these challenges when building the company’s autonomous identity security agent Dorian, which detects and mitigates threats across different digital identity providers.

“Without proper identity infrastructure, you can build an interesting AI agent — but you can’t productize it, scale it, or sell it,” points out Aviad Mizrachi, co-founder and CTO of Frontegg.

The new Frontegg.ai gives developers an identity layer with a multi-tenant onboarding flow, enterprise-grade policy controls, token rotation and security controls to support both B2B and B2C agent applications.

Anetac ups identity visibility

Silicon Valley-based Anetac has launched the new Human Link Pro software to provide real-time visibility into human and non-human identities across chains. This visibility can reveal privilege escalations and inheritance risk often hidden from traditional solutions, the company says.

The feature is built into Anetac’s Identity Vulnerability Management Platform. The enhancement is already used financial services, healthcare, retail and critical infrastructure organizations to detect and respond to vulnerabilities across all identity types.

Human Link Pro’s major features are detection of behavioral anomalies, compromised credentials, joiner-mover-leaver (JML) gaps, unconstrained delegation and dormant or orphaned accounts.

Anetac demonstrated the new capabilities at RSA Conference 2025 last week.

Article Topics

AI agents  |  Anetac  |  biometric authentication  |  biometric binding  |  biometrics  |  digital identity  |  Frontegg  |  identity access management (IAM)  |  identity verification  |  ZeroBiometrics