Artificial intelligence (AI) is reshaping medicine, promising advances in diagnosis, monitoring, and treatment, and psychiatry will be no exception. Yet the field remains fragmented: ethical guidelines, technical standards, and clinical workflows have evolved in parallel, creating uncertainty about how to integrate AI safely and meaningfully into psychiatric care. Existing frameworks often address isolated domains (explainability, data protection, or harm prevention [HP]) without providing a coherent structure that connects them to everyday clinical realities. This article introduces a global framework for the responsible integration of AI in psychiatry, built on 4 non-negotiable system capabilities: Explainable AI to ensure transparency and trust; Shared Decision-Making to protect patient autonomy; Electronic Health Record integration to secure continuity and accountability; and HP to embed multilayered safety controls. Together, these pillars define a responsibility-by-design approach that aligns technological development with psychiatry's ethical foundations. The framework offers clinicians, policymakers, and developers a roadmap for aligning innovation with human values and measurable improvements in clinical outcomes. By translating ethical commitments into auditable, non-negotiable system capabilities, it establishes a concrete foundation for regulatory oversight, guideline endorsement, and responsible AI deployment in psychiatry.
原文链接
分享
引用
