Permissions

Permissions, in the context of AI systems, refers to the rules that determine what an AI model, agent, or connected tool is allowed to read, change, or trigger inside an organization's systems, such as files, email, customer records, or payment functions. Weak or overly broad permissions are a common source of AI incidents, because an AI agent given more access than its task requires can expose sensitive data or take actions no one approved. For governance teams, reviewing and limiting these permissions is a practical control that supports least-privilege access, auditability, and accountability for automated decisions.

1 item