Answers are grounded in this paper and cite their sources.
Generative AI is changing not only how content is created on digital content platforms but also who provides the AI used to create it. Platforms may offer generative AI using data pooled across creators, whereas creators may provide AI services based on their own content. Understanding how these provision regimes affect creators’ investment incentives and the value generated by AI is therefore important. We develop a game-theoretic model in which human-generated content competes with AI-generated content while contributing to AI learning. The results show that platform provision can induce higher human-generated content quality than creator provision when AI learning effectiveness lies within a certain range and the platform’s pooled-data advantage is limited. A stronger pooled-data advantage can make creator provision more favorable for sustaining human-generated content quality. Platform provision can nevertheless generate higher AI-generated content quality even when creator provision induces higher human-generated content quality, provided that the pooled-data advantage is sufficiently strong. We further find that AI value sharing can encourage or discourage quality investment under platform provision, while its effects on regime preferences and preference alignment depend on market conditions. Thus, the provision regime preferred by the creator need not be the one that sustains higher human-generated content quality.
No takes yet. Share an insight, caveat, or question.
Xue‐Yan Wu (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: