Responsible AI in Public Health Systems: Governance Principles and an Implementation Model for Indian States

Artificial intelligence (AI) is entering Indian public health systems through screening tools, clinical decision support, disease surveillance, health-record analytics and citizen-facing chatbots. State health departments, which run most public facilities, are often the buyers and operators of these tools, yet they have limited guidance on how to evaluate, procure, deploy and monitor them responsibly. International and national guidance — including the World Health Organization's guidance on the ethics and governance of AI for health and the Indian Council of Medical Research's ethical guidelines for AI in biomedical research and healthcare — sets out principles, but translating them into state-level practice remains a gap. This paper proposes an implementation model for responsible AI in state public health systems, organised around the AI life cycle: need and problem definition; data governance; evaluation and validation; procurement; deployment and workflow integration; monitoring and incident management; and decommissioning. It pairs each stage with governance principles, minimum requirements and responsible roles, and recommends an institutional structure — a state AI-in-health committee supported by a technical assessment cell. The paper draws on published guidance and research and on the author's experience in state digital health programmes; it is a policy and implementation framework, not an empirical study.