The Creative Divide: How AI Shapes Value and Inequality Among Creators
36 Pages Posted: 7 May 2025
Date Written: September 16, 2024
Abstract
The rapid development and widespread adoption of artificial intelligence (AI) are transforming many industries, including creative arts and software—fields traditionally regarded as strongholds of human creativity. A central question that remains debated is whether AI will democratize creative production by narrowing the output gap between high-skilled and low-skilled creators. This study investigates how AI adds value to the creative process and how it influences the product-value gap across skill levels. We model creative products as having two key attributes: ideation and execution. AI affects both through two distinct mechanisms: automation, which substitutes for certain execution tasks, and augmentation, which enhances creators’ ideation capabilities. We find that AI consistently improves product value for creators at all skill levels. However, the magnitude of this improvement varies across creators: AI tends to widen the product-value gap when the high-skilled creator is ideation savvy (i.e., holds a greater advantage in ideation than execution) or when the high-skilled creator is execution savvy but the automation is relatively weak. In other scenarios, AI may instead narrow the gap. These outcomes are driven by shifts in effort allocation, as AI both reduces the need for human effort in execution and increases the effectiveness of ideation. As a result, creators reallocate effort from execution to ideation. The widening of the gap may result from growing disparities in both attributes—or from ideation alone—or from increased complementarity between ideation and execution. Finally, we show how the impact of AI on the product-value gap is moderated by the market’s relative appreciation for ideation versus execution, as well as by the strength of AI’s automation and augmentation capabilities.
Keywords: AI, Automation and augmentation, Creative product, Analytical modeling
Suggested Citation: Suggested Citation