Governing AI in Pharmaceutical Operations: How to decide what AI may do and whether people can use it reliably

AI should not support pharmaceutical quality work until the organization has defined the task, established what the system may do, and tested whether the complete human and AI workflow can perform representative work reliably. This working paper presents the AI-Supported Work Authorization Cycle, six gated stages connecting use inventory and ownership, risk and intended use, authority and records, workflow testing, routine authorization and monitoring, and change, correction, reauthorization, or retirement. It proposes an evidence-centered operating method for deciding whether AI-supported work may proceed, under what limits, and what finding should cause the organization to restrict or stop it. Supporting templates are available at https://www.gxpframe.com/governing-ai/.