AI Governance
Also known as: responsible AI, AI risk management, model governance
AI governance is the set of policies, processes, and controls organizations use to develop, deploy, and operate AI systems responsibly — covering risk management, model lifecycle, accountability, data protection, and regulatory compliance.
Detailed explanation
AI governance translates principles like fairness, accountability, transparency, and safety into operational practice. It typically covers model risk policies, approval workflows, model and data inventories, evaluation requirements, incident response, audit logging, and alignment with regulations (EU AI Act, India’s emerging AI framework, sector-specific guidance).
Practical AI governance focuses on the system, not just the model: how data is sourced and consented, how prompts and outputs are logged, who can deploy what, how changes are reviewed, and how harms are detected and remediated. For LLM systems, additional controls address prompt injection, data leakage, hallucination risk, and dependency on third-party models.
Mature programs integrate governance into the development lifecycle (governance-by-design) rather than bolt it on as a gating function. Documentation artifacts include model cards, eval reports, data sheets, and impact assessments.