The digital identity industry has been diligently working on how to authenticate people without risking exposure to fraud attacks, particularly phishing, even as a new problem has emerged. AI agents are emerging as a primary concern for enterprise cybersecurity, Hypr data shows. Yet Indicio starts its new paper on agentic identity with a series of reports that forecast trillions of dollars in agentic commerce over the next half-decade.<\/p>\n
Identity is a necessary component of agentic governance<\/a>, and cryptography<\/a> is the mechanism for secure verification. These are already familiar foundations<\/a> for identity and access management (IAM). But the authorization that extends from that identity layer must go a step further, and that remains a heavy lift, at least for now.<\/p>\n Generative AI and agentic AI have surpassed stolen credentials as the top concern among identity security professionals, according to Hypr<\/a>\u2019s latest State of Passwordless Identity Assurance<\/a> report.<\/p>\n Automation has changed the scale of the identity security battle, Hypr says, and rendered point-in-time security measures inadequate. This view aligns with Google Threat Intelligence Group\u2019s recent observation that agentic AI is being used to automate credential theft<\/a>. Identity verification is the step enterprises need to keep up with agentic identity security requirements, Hypr suggests.<\/p>\n \u2018You can\u2019t prove I lost your credentials, or even that I\u2019m a robot\u2019<\/p>\n API keys and OAuth tokens<\/a> are vulnerable to credential theft and do not include information about the context they are meant for. That makes AI agents deployed by enterprise infrastructure with these tools potentially insecure and dangerous.<\/p>\n At the same time, the humans are continuing to resist the measures that would make their credentials more secure. The adoption of passwordless authentication has hit a plateau, according to Hypr\u2019s data, even as agents surpass people in password leakage.<\/p>\n Even confirming whether a password or other credential\u2019s compromise is a person or agent\u2019s fault may be getting harder.<\/p>\n OpenAI Member of Technical Staff Sharif Shameem shared a video on X<\/a> of GPT-6 Astra passing all 48 levels of the \u201cI\u2019m Not a Robot<\/a>\u201d game created by Neal Agarwal. The game is based on the Captchas that still differentiate people from bots on most of the net.<\/p>\n Decentralized ID and VCs to the rescue<\/p>\n The situation is far from hopeless, however.<\/p>\n Indicio<\/a>\u2019s new white paper on \u201cIdentity, Commerce, and the Agentic Web<\/a>\u201d argues that decentralized identity and verifiable credentials represent the way to deliver the kind of secure identity verification Hypr refers to.<\/p>\n The ecosystem is maturing with protocols to allow agents to interact with systems, like MCP, and each other, like Google\u2019s A2A. But three conditions need to be satisfied for those interactions to be trusted.<\/p>\n Agents need cryptographic binding to a legitimate person or organization, authorization for a given action and the consent of the party they are performing it on behalf of for secure compliance, Indicio says. The decentralized digital identity models that powered web3 are also applicable to AI agents, with VCs providing a way to verify the agent, their instructions and who they are bound to.<\/p>\n Investment is already shifting in recognition of the impact of agents on commerce, and trust on agentic commerce, according to Indicio. It will go to infrastructure to support the governance of autonomous software agents, along with the verifiable agent credentials themselves.<\/p>\n The pricing for the technologies that provide digital trust will also have to evolve. IAM has long relied on a \u201cper-seat\u201d pricing model, where the seats each represent a human user. As those users represent a declining share, usage-based models make more sense.<\/p>\n Luckily, the analysts say agentic commerce will be generating plenty of money to pay for any reasonable pricing model.<\/p>\n \t\t\t\tArticle Topics<\/p>\n agentic commerce<\/a> \u00a0|\u00a0 AI agents<\/a> \u00a0|\u00a0 digital identity<\/a> \u00a0|\u00a0 HYPR<\/a> \u00a0|\u00a0 identity access management (IAM)<\/a> \u00a0|\u00a0 Indicio<\/a><\/p>\n \t\t\t\t\tLatest Biometrics News<\/p>\n \t\t\t\t \n\t\t\t\t\t\t\t\tSep 10, 2026, 4:52 pm EDT\t\t\t\t\t\t\t<\/p>\n Clearview AI has developed an experimental AI tool that could extend a facial recognition search into broader automated research on\u2026<\/p>\n \t\t\t\t \n\t\t\t\t\t\t\t\tSep 10, 2026, 11:35 am EDT\t\t\t\t\t\t\t<\/p>\n Africa\u2019s next wave of digital identity investment may have surprisingly little to do with biometrics, wallets or credentials. In early\u2026<\/p>\n \t\t\t\t \n\t\t\t\t\t\t\t\tSep 10, 2026, 10:59 am EDT\t\t\t\t\t\t\t<\/p>\n Europe\u2019s digital identity ambitions are approaching an important deadline. By the end of 2026, EU Member States are expected to\u2026<\/p>\n \t\t\t\t \n\t\t\t\t\t\t\t\tSep 10, 2026, 10:55 am EDT\t\t\t\t\t\t\t<\/p>\n By Joshua W.J. Brown Biometric authentication has mature ways to talk about comparison error. That is exactly why accessibility failures\u2026<\/p>\n
\n\t\t\t\t\t\u00a0\t\t\t\t<\/a>\t\t\t\t<\/p>\n
\n\t\t\t\t\t\u00a0\t\t\t\t<\/a>\t\t\t\t<\/p>\n
\n\t\t\t\t\t\u00a0\t\t\t\t<\/a>\t\t\t\t<\/p>\n
\n\t\t\t\t\t\u00a0\t\t\t\t<\/a>\t\t\t\t<\/p>\n