Teaching Pharmacists to Challenge Artificial Intelligence: A Critical Review of Curricula, Competency Claims, and Assessment

Artificial intelligence is entering pharmacy education faster than evidence is establishing what pharmacists must know, demonstrate, and resist. This critical review examined whether curricula, competency claims, and assessments prepare pharmacists to challenge artificial intelligence outputs rather than merely operate tools. A transparent critical-review method with SANRA-informed quality controls was applied to peer-reviewed literature published from 2017 through 2025. Searches of PubMed/MEDLINE, targeted journal and publisher platforms, and backward and forward citation trails were executed on 2 August 2026 with publication-date censoring. Eligible reports addressed pharmacy curricula, competency frameworks, educational interventions, assessment, or directly transferable health-professions evidence relevant to verification, uncertainty, bias, hallucination, refusal, escalation, accountability, workplace performance, or patient communication. Records were deduplicated, screened, linked to study families, extracted using predefined fields, and appraised according to design. Heterogeneous findings were compared narratively without quantitative pooling. Fifty-nine retrievals yielded 38 unique records, 27 assessed full texts, and 18 included reports representing 18 studies: six pharmacy-specific and 12 transferable reports. Pharmacy-specific evidence chiefly described exposure, perceptions, task completion, item generation, or artificial-intelligence-assisted grading. Direct assessment of independent verification, justified refusal, escalation, documented accountability, or workplace transfer was scarce. Transferable frameworks widened the competency scope but did not validate pharmacy-specific performance. The review was limited by a single-reviewer workflow, source-interface constraints, methodological heterogeneity, and the later search execution date. Pharmacy education should distinguish tool use from challenge competence and align claims with observable, authentic performance across realistic medication decisions involving patients, teams, and governance constraints. Proposed domains require prospective validation before curricular standardization.