现代产业经济学系列讲座第302期

时    间：2026年9月11日13:30-15:00

地    点： 复旦管院史带楼303室

题    目：Generative AI as a Competitor: Content Creation and Platform Strategy

主讲人：陈丽梅  华南师范大学经济与管理学院讲师

主持人：李玲芳 教授

论文摘要：With advances in generative AI, content platforms increasingly use creators’ output as training data to supply their own AI-generated content, which can compete with those same creators for consumer attention. This paper therefore studies how a content platform manages human-content production under this tension by adjusting its revenue-sharing rule and choosing its AI strategy. We develop a game-theoretic model in which creators allocate investment across content dimensions in response to revenue sharing and AI competition. We obtain three main findings. First, the revenue-sharing rule shapes both creators’investment incentives and the competitive role of AI content, making the optimal rule regime-dependent. At an intermediate capability level, the platform discretely lowers sharing to use AI as a binding entry threat, then switches to a qualitatively different policy once direct AI competition becomes profitable. Second, platform-supplied AI distorts creators’ investment mix toward less productive content attributes that AI cannot learn to reproduce, but it need not reduce human-content quality. Under optimal sharing, quality is discontinuous and nonmonotonic in AI capability and may be lowest at intermediate levels rather than when AI is strongest. Finally, although supplying AI content is always profitable for the platform, it may restrain AI capability to protect the human-content base it continues to monetize, even in the absence of AI development costs. At the platform’s optimal capability, however, AI still diverts consumer attention from human content and leaves creators worse off than without AI, showing that protecting content quality does not amount to protecting creators.

主讲人介绍：陈丽梅，华南师范大学经济与管理学院青年英才，复旦大学经济学博士，研究领域为产业组织理论、契约理论、数字经济等，具体包括平台激励、平台自我优待、平台数据使用等问题。目前研究成果发表在重要的国际期刊上，并获得了2026年度的国家自然科学基金青年（C类）项目。