AI Governance in Practice

Develop a framework for regulating AI use in public institutions and law enforcement. What should be tightly controlled, what should require human oversight, and what can be left to agencies or markets to experiment with? Discuss the core concerns: bias, accountability, surveillance, data protection, explainability, auditability, and procurement standards. Then compare a rules-first approach with a sandbox approach, and recommend a model that could scale as AI tools evolve.