Identity Governance: The New Challenge in AI Security, According to Hush Security

Israeli cybersecurity startup Hush Security has quickly made an impact in the non-human identity security space, securing a $30 million Series A round led by Battery Ventures, YL Ventures, and Akamai Technologies. The company believes that the conversation around enterprise AI security has shifted, with organizations now focusing on deploying autonomous software agents that require a new security model.

CEO and co-founder Micha Rave notes that the company’s original focus on securing non-human identities has evolved to address the challenge of safely allowing AI agents to operate within production systems. This shift was prompted by a high-profile hacking incident involving an autonomous AI agent at Hugging Face. With the projected increase in AI agents running within Fortune 500 organizations, Hush emphasizes the importance of identity as the critical control point for enterprise AI.

Hush’s Identity Gateway platform extends its existing non-human identity solution to provide governance for AI agents. By assigning each agent its own identity, associating it with a human owner, and enforcing least privilege access, organizations can ensure that actions are logged, attributed, and revoked as needed. This approach addresses the identity challenges posed by autonomous AI agents, which often operate across multiple enterprise systems with broad permissions.

The company highlights the need for dedicated governance for AI agents, which represent a third category of identity alongside human and machine identities. Hush’s platform acts as a bridge between traditional identity providers and secrets managers, offering runtime visibility, risk analysis, and identity-based access controls for AI agents and non-human identities. This approach has attracted customers like Kyndryl, who have deployed Hush internally and are offering the platform to enterprise clients.

As enterprises increasingly deploy AI agents for various tasks, the focus of AI security is shifting from model vulnerabilities to operational questions around access control and governance. Hush’s identity-centric approach positions identity as the enforcement layer for managing the identities of autonomous software agents securely. The company’s participation in the broader AI security market reflects a recognition that securing AI models is just one part of the equation, with identity management playing a crucial role in ensuring the security of enterprise AI initiatives.