Competence before Confidence in Pharmacist Use of Generative Artificial Intelligence

Generative artificial intelligence is entering pharmacy through medication-information searches, evidence synthesis, documentation, patient communication, clinical reasoning support, safety surveillance, education, and administrative work. Its fluent output can encourage confidence before users have demonstrated the ability to define an appropriate task, verify evidence, recognize errors, communicate uncertainty, and retain professional accountability. This article proposes an original Pharmacist Generative-AI Competency Framework that treats competence as task-specific, observable, context-dependent, and continuously requalifiable. The framework connects authorized pharmacy tasks with nine domains: task framing and authorization; model and data literacy; evidence retrieval and provenance; verification and error recognition; medication safety and uncertainty; human–AI and patient communication; workflow accountability; equity and representation; and monitoring and requalification. Four proposed levels extend from supervised constrained use to critical governance, with advancement controlled by case-based assessment gates rather than self-reported confidence or course completion. The framework distinguishes model performance, pharmacist task performance, pharmacist–AI team performance, workflow consequence, and patient-relevant outcome. It also places individual competence inside organizational conditions such as evidence access, privacy protection, supervision, auditability, incident response, and change control. Validation would require reliable scoring, external assessment across practice settings, medication-safety testing, human-factors evaluation, equity analysis, and longitudinal evidence after model or workflow change. The framework is non-empirical and does not establish clinical benefit, authorization, regulatory acceptance, or deployment readiness. Its contribution is a testable architecture for defining what pharmacists should be able to demonstrate before confidence in generative-AI use is treated as professionally meaningful.