Abstract: Generative Artificial Intelligence (GAI) has restructured the entire logic of the advertising industry. The hyper-realistic and highly concealed false advertisements it generates challenge the foundational presumptions of traditional regulation—namely "identifiable subjects, verifiable facts, controllable harms, and post-accountability." This disruption triggers systemic dilemmas, including broken attribution chains, failure in fact-finding, distorted review mechanisms, and blocked remedial channels. The solution lies in transitioning the regulatory paradigm from "individual behavior control" to "systemic risk governance." The core strategy involves reconstructing platform liability to transform platforms from passive intermediaries into "gradient gatekeepers, " establishing a new governance logic centered on risk tracing, source control, and process intervention. Accordingly, a comprehensive regulatory system covering ex-ante prevention, in-process intervention, and ex-post accountability is proposed. Ex-ante measures focus on mandatory labeling and trusted traceability to ensure procedural compliance; in-process measures employ algorithmic transparency and dynamic risk interception mechanisms for preemptive risk interception; ex-post measures utilize "presumption of knowledge" and "substantial participation" doctrines to solidify platform responsibility. Complemented by regulatory technology and multi-stakeholder collaboration, this study provides theoretical support and institutional solutions for updating digital-era advertising laws.