Trust in AI systems cannot be built on promises alone – it requires rigorous testing throughout the entire lifecycle of these systems. True assurance means systematically examining models and agents for hallucinations, safety failures, data leakage, and potential misuse. For testing to be credible, it must be context-specific, transparent, independently verifiable, and regularly updated. AI assurance represents a practical, scalable mechanism for earning trust rather than merely a marketing slogan. Hear from Lee Wan Sie, Cluster Director for AI Governance at Singapore's Infocomm Media Development Authority, on how systematic testing frameworks can build genuine trust in AI systems.