Ethical Implications of AI-Based Decision Support in Dispensing High-Risk Medications

Abstract

Artificial intelligence (AI)-based decision support systems (DSS) are increasingly embedded in pharmacy workflows to flag contraindications, calculate individualized doses, and prioritize clinical interventions, especially for high-risk medications such as anticoagulants, chemotherapeutic agents, insulin, and opioids.While these tools promise reductions in adverse drug events (ADEs) and improved operational efficiency, they simultaneously raise complex ethical questions around autonomy, accountability, bias, transparency, data governance, and professional identity.This 7,000-word manuscript examines the ethical implications of deploying AI-DSS in the dispensing of high-risk medications.Using a mixed-method design (a national survey of 312 pharmacists and qualitative interviews with 22 stakeholders), it interrogates how clinicians perceive and negotiate algorithmic recommendations, where liability is located when harm occurs, and how design choices influence equity in medication safety.Results show that pharmacists value AI alerts for rare but catastrophic errors, yet worry about alert fatigue, opaque logic, and systemic bias against marginalized populations.Respondents advocated for shared accountability models, context-aware explainability, and participatory governance structures.The study concludes with a framework-PASTOR (Proportionality, Accountability, Safety/Beneficence, Transparency/Traceability, Oversight/Justice, Respect for Autonomy)-to guide ethical implementation.Recommendations include layered explainability, harm-register audits, and continuous socio-technical evaluation.The paper underscores that ethical AI in medication dispensing is a process, not a product, requiring ongoing reflexivity, stakeholder inclusion, and regulatory agility.

Ähnliche Arbeiten