Google has introduced Beyond Zero, a security framework designed for the AI era that extends Zero Trust principles to autonomous AI agents. The new approach shifts access decisions from the application level to individual resources and actions, merging static authorization controls with dynamic AI-driven decisions to enable machine-speed enforcement for both humans and agents. Beyond Zero represents what Google calls a necessary evolution from its 2014 BeyondCorp model, which the company says no longer fits an enterprise environment where millions of AI agents operate alongside human workers.
The Beyond Zero model operates on five core principles: authorization at the level of individual actions and resources across interfaces and APIs; a blend of static policies and dynamic controls for higher-risk scenarios; automatically enriched context about users, actions, data, and risks; automated investigation triggered by risk signals; and challenges or containment measures that can require additional verification or telemetry from users and AI agents. Google says deployments are currently internal-only, the components aren't all built yet, and more papers are forthcoming. The company published its original BeyondCorp whitepaper in 2014, outlining its strategy for replacing network-perimeter security with a zero-trust model for enterprise access.
Joseph Valente, formerly director of product management at Google, and Michal Zalewski, distinguished security researcher and formerly at Google, write in the paper: "The assumptions underpinning BeyondCorp—that accessors are human, that actions occur at human speed, and that applications are the correct boundary for trust—are no longer sufficient." Heather Adkins, VP of security engineering at Google, and Archana Ramamoorthy, senior director at Google, describe the enterprise landscape as entering a new era shaped by artificial intelligence, where AI is transforming the assumptions around how enterprise security works. Valente explains on LinkedIn that continuous authorization of every action at scale initially seemed excessive, but became necessary as hundreds of thousands of workers were joined by millions of agents.
To adopt the latest model, SaaS vendors will need to expose action-level authorization, standards need to mature, and smaller security teams face difficult issues around false positives, intent, auditing, and cost. Kane Narraway, security manager at Canva, is positive about Beyond Zero's goals but skeptical about how quickly ordinary enterprises can implement them, warning that the paper is aspirational and follows the same playbook as 2014: publish the vision, describe the architecture at a high level, and encourage the industry to build towards it collaboratively. Community reaction on Hacker News has been largely skeptical, with commenters questioning the reliability and complexity of AI-driven authorization. Some practitioners argue that access control is traditionally a hard security boundary, whereas probabilistic decisions are inherently harder to predict, test, and audit. Firas Durri writes that preparing for AI agents means expanding auditability and reversibility in software.
Google presents Beyond Zero as an ongoing effort and says it plans to publish further details on its implementation and operational considerations, but no specific deadlines have been published. The company frames the model as essential for environments where AI agents are deployed globally to increase operational velocity and boost productivity. The shift represents a fundamental rethinking of enterprise security boundaries, moving from application-level trust to granular, continuous authorization of every individual action. Organizations that implement Beyond Zero will need to balance the promise of machine-speed security enforcement against the practical challenges of debugging non-deterministic access decisions and managing false positives at scale. The tension between Google's aspirational vision and the readiness of existing security tooling may determine whether Beyond Zero becomes an industry standard or remains confined to tech giants with the resources to build custom infrastructure.