Abstract: As a disruptive technology, generative artificial intelligence (GAI) drives industrial transformation and technological advancement while simultaneously posing systemic challenges to existing industrial ecosystems, governance paradigms, and social order. This intensifies the core tension between technological development and risk regulation, termed the "Collingridge Dilemma". This study addresses this critical governance challenge by systematically reviewing and critically evaluating existing regulatory approaches and mainstream international models. The analysis reveals structural limitations across these frameworks, including misalignment between regulatory cycles and technological iterations, the failure of traditional tools to anticipate emerging risks, and value conflicts between safety precautions and innovation tolerance. To address this dilemma, the study anchors itself in China's principle of "inclusive and prudent regulation". After elucidating its theoretical foundations and practical implications, the paper constructs a systematic regulatory framework comprising four integrated mechanisms: duty-clarification mechanism, experimentation and error-tolerance mechanism, multi-stakeholder participation mechanism, and oversight and accountability mechanism. This framework transcends the binary opposition of strict control versus laissez-faire, offering an institutional innovation pathway for maximizing GAI's compliant development within legal and ethical boundaries, thereby contributing Chinese solutions to global governance.