Abstract: The rapid advancement of generative AI has significantly enhanced content creation capabilities for marketing practitioners. However, influencer marketing— a crucial engine for stimulating domestic demand and boosting consumption — now faces practical challenges of low sales conversion rates due to content homogenization. This study proposes that leveraging generative AI to create sponsored content adhering to the optimal distinctiveness principle may resolve this problem. Based on social media affordance theory and optimal distinctiveness theory, the research innovatively employs a mixed-methods approach integrating multimodal machine learning, quantitative marketing analysis, and behavioral experiments. It aims to: (1) identify feature dimensions of multimodal sponsored content and measure sponsored content similarity; (2) unveil the mechanisms of sponsored content similarity on influencer marketing effectiveness to identify optimal distinctiveness thresholds; (3) explore technical pathways for generating optimally distinctive multimodal sponsored content using generative AI. The findings will establish a comprehensive theoretical framework for analyzing influencer marketing effectiveness while providing actionable scientific solutions to practical challenges. The outcomes not only align with the core demands of developing an "AI+Consumption" ecosystem in the digital economy era but also demonstrate significant practical significance for expanding domestic demand and revitalizing consumer markets.

Key words: influencer marketing, consumer psychology, multi-modal analysis, generative AI, optimal distinctiveness theory