Data Sharing with a Generative AI Competitor

TL;DR

This paper models the economics of data sharing between a competing content firm and a Generative AI platform as a two-stage Stackelberg game, where the Firm first selects a share level $\alpha$ and the GenAI platform then decides how much external data $x$ to acquire, under an exogenous per-unit data price $m$ and a bounded data market. Using a bi-linear traffic function, the authors derive a complete subgame-perfect equilibrium, revealing two possible equilibria: either the Firm shares enough data so GenAI relies on Firm up to $\alpha=R_G$ and buys no external data ($x=0$), or the Firm shares $R_F$ and GenAI purchases all external data ($x=1-R_F$). They show that costly data sharing can arise (the Firm may pay to share data), establish conditions for Pareto-improving data-sharing prices, and analyze how pricing can be tuned to favor additional Firm data sharing, more expert data acquisition, or a balance of both. The work offers practical guidance for platforms and policymakers on designing data exchange mechanisms in the GenAI era, including how to structure incentives and regulation to promote efficient data sharing. Mathematical extensions generalize the results to traffic functions with data overlap, reinforcing the robustness of the key insights across plausible market dynamics.

Abstract

As GenAI platforms grow, their dependence on content from competing providers, combined with access to alternative data sources, creates new challenges for data-sharing decisions. In this paper, we provide a model of data sharing between a content creation firm and a GenAI platform that can also acquire content from third-party experts. The interaction is modeled as a Stackelberg game: the firm first decides how much of its proprietary dataset to share with GenAI, and GenAI subsequently determines how much additional data to acquire from external experts. Their utilities depend on user traffic, monetary transfers, and the cost of acquiring additional data from external experts. We characterize the unique subgame perfect equilibrium of the game and uncover a surprising phenomenon: The firm may be willing to pay GenAI to share the firm's own data, leading to a costly data-sharing equilibrium. We further characterize the set of Pareto improving data prices, and show that such improvements occur only when the firm pays to share data. Finally, we study how the price can be set to optimize different design objectives, such as promoting firm data sharing, expert data acquisition, or a balance of both. Our results shed light on the economic forces shaping data-sharing partnerships in the age of GenAI, and provide guidance for platforms, regulators and policymakers seeking to design effective data exchange mechanisms.

Figures (3)

• Figure 1: The utility of Firm (given the best reply of GenAI) as a function of its data sharing level $\alpha$.
• Figure 2: An illustration of the set of Pareto improving prices corresponding to Example \ref{['example-1']}. Notice that this example falls into the second case of Proposition \ref{['cor:pareto']}, as the set of prices splits into two disjoint (non-positive) intervals.
• Figure 3: Sensitivity analysis for $r^f, r^g, c, m$. The top row (figures a–c) varies $r^f$ and $r^g$: figures a and b show the utilities of the Firm and GenAI, respectively, while figure c presents the induced equilibrium for each parameter combination. The bottom row (figures d–f) varies $c$ and $m$: figures d and e describe the corresponding utilities, and figure f shows the resulting equilibrium.

Theorems & Definitions (16)

• Theorem 1: informal version of Theorem \ref{['thrm:eq-analysis']}
• Example 1
• Theorem 2
• Proposition 1
• Definition 1
• Proposition 2
• Proposition 3
• proof : Proof of \ref{['lemma: condition1']}
• Proposition 4
• proof