As interagency intelligence systems grow in scale and interconnectivity, Artificial Intelligence (AI) has moved from a supporting role to a central mechanism in how democratic states manage public security and address risk. By enabling near-instant information flows between domains as varied as law enforcement, border management, and public health, AI can heighten both coordination and responsiveness. Yet the same capabilities that allow rapid action also sharpen long-standing concerns over privacy, legality, and the resilience of democratic oversight. This chapter explores how AI shapes the governance of sensitive information through three key functions - access filtering, data aggregation, and dispersion control. Using case studies from both the EU and the United States, it illustrates how each function generates operational value while carrying distinct vulnerabilities: bottlenecks that can slow urgent interventions, profiling capacities that invite misuse, and monitoring practices that may edge into overreach. To meet these challenges, a hybrid governance model is proposed, one that weaves together deontological safeguards to set firm limits, utilitarian reasoning to calibrate flexibility, and care-based responsiveness to adapt to context and preserve trust. This ethical architecture is embedded within a converged legal framework—encompassing the GDPR, the Law Enforcement Directive, U.S. sectoral statutes, and international instruments such as the OECD AI Principles—and is made operational through five design requirements: selective filtering, explainable logic, distributed oversight, correction mechanisms, and ethics by design. By treating proportionality, necessity, and transparency not as slogans but as operational rules, AI-enabled intelligence systems can better reconcile institutional effectiveness with the protection of rights—ensuring that the algorithmic infrastructures of tomorrow reinforce, rather than corrode, the foundations of democratic governance.