Summary
Establishes a Ministry-wide framework for the safe use of AI language tools and large language models in healthcare. AI outputs must remain supplementary to professional judgment; identifiable or confidential patient data may not be entered into unapproved tools; uses are classified into three risk levels; and AI may not be the sole basis for diagnosis, treatment, prescribing, dosage calculations, test interpretation, admission or discharge, referrals, triage, or disease-severity assessment. Patient-facing AI tools require formal Ministry evaluation and accreditation.
Healthcare Implications
Requires healthcare personnel and facilities to verify AI outputs, preserve human clinical responsibility, use only approved systems for patient or operational data, review AI-generated medical-record content before inclusion, report privacy or safety incidents, and obtain formal assessment before deploying patient-facing AI services.