AI agents in infrastructure: new permissions, new dependencies and new operational risk

info@teralevel.com (TeraLevel) — Tue, 28 Apr 2026 10:00:00 +0000

AI agents do not only automate tasks: they also introduce new identities, permissions and dependencies within infrastructure.

ComputerWeekly recently analysed how the adoption of AI agents is growing faster than many organisations’ ability to secure this new operating model. The discussion is no longer limited to generative AI as a support tool, but to systems capable of acting on applications, repositories, APIs, tickets, cloud platforms and internal workflows.

Source: ComputerWeekly – AI agents are here. Are we ready for the security implications?

IAM on TeraLevel

AI agents in infrastructure: new permissions, new dependencies and new operational risk

AI agents do not only automate tasks: they also introduce new identities, permissions and dependencies within infrastructure.