AI regulation is widely discussed as if it were about controlling models. That framing is convenient, but wrong. The dominant regulatory exposure does not arise from how AI systems are built. It arises from how AI-generated statements are relied upon in decisions that carry legal, financial, or reputational consequences. Across jurisdictions, regulators are converging on a single expectation: If an AI-generated statement influences a consequential decision, the organization that relied on it must be able to reconstruct what was said, when, and in what context. This article separates fact from fiction in current AI regulation and maps enforceable obligations to a specific and under-governed risk surface: AI Reliance.