When capability hurts: Strategic adoption and disclosure of artificial intelligence in creator competition

Generative artificial intelligence (AI) reshapes digital content platforms by lowering barriers for creators to produce content, while widespread algorithm aversion among consumers creates uncertainty about its impact on stakeholders. Whereas prior literature focuses on monopolistic decision-making, our framework shows how competition fundamentally reshapes disclosure incentives by linking consumer heterogeneity to price interaction among creators. To examine this tension, we develop a game-theoretic framework of two competing creators differing in creativity and heterogeneous consumers who differ in both their perceived content quality and ability to recognize AI adoption. In this setting, creators may adopt AI to enhance content quality and must decide whether to disclose its use. The analysis shows that equilibrium outcomes are jointly affected by the AI capability, algorithm aversion, and opacity penalty. Although AI adoption consistently reduces rival profits, greater AI capability can intensify price competition to the extent that AI adoption becomes unprofitable for the weaker creator. To avoid such cases, the weaker creator rejects to use highly capable AI. When AI capability is relatively low, a moderate increase in the proportion of sophisticated consumers makes the weaker creator more likely to disclose AI adoption, while it declines AI usage if sophisticated consumers account for a significant share of the market. These findings suggest that more capable AI does not necessarily benefit either creators or consumers, as its consequences hinge on consumer heterogeneity and competitive dynamics. Accordingly, AI policymakers should weigh transparency requirements against the risk of discouraging adoption, since mandatory disclosure may backfire in markets with strong algorithm aversion. © 2026 Elsevier Ltd.