Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

When AI Levels the Playing Field: Skill Homogenization, Asset Concentration, and Two Regimes of Inequality

Xupeng Chen, Shuchen Meng

TL;DR

This work formalizes this tension in a task-based model with endogenous education, employer screening, and heterogeneous firms, and yields two regimes whose boundary depends on AI's technology structure and labor market institutions (rent-sharing elasticity, asset concentration).

Abstract

Generative AI compresses within-task skill differences while shifting economic value toward concentrated complementary assets, creating an apparent paradox: the technology that equalizes individual performance may widen aggregate inequality. We formalize this tension in a task-based model with endogenous education, employer screening, and heterogeneous firms. The model yields two regimes whose boundary depends on AI's technology structure (proprietary vs. commodity) and labor market institutions (rent-sharing elasticity, asset concentration). A scenario analysis via Method of Simulated Moments, matching six empirical targets, disciplines the model's quantitative magnitudes; a sensitivity decomposition reveals that the five non-$Δ$Gini moments identify mechanism rates but not the aggregate sign, which at the calibrated parameters is pinned by $m_6$ and $ξ$, while AI's technology structure ($η_1$ vs. $η_0$) independently crosses the boundary. The contribution is the mechanism -- not a verdict on the sign. Occupation-level regressions using BLS OEWS data (2019--2023) illustrate why such data cannot test the model's task-level predictions. The predictions are testable with within-occupation, within-task panel data that do not yet exist at scale.

When AI Levels the Playing Field: Skill Homogenization, Asset Concentration, and Two Regimes of Inequality

TL;DR

This work formalizes this tension in a task-based model with endogenous education, employer screening, and heterogeneous firms, and yields two regimes whose boundary depends on AI's technology structure and labor market institutions (rent-sharing elasticity, asset concentration).

Abstract

Generative AI compresses within-task skill differences while shifting economic value toward concentrated complementary assets, creating an apparent paradox: the technology that equalizes individual performance may widen aggregate inequality. We formalize this tension in a task-based model with endogenous education, employer screening, and heterogeneous firms. The model yields two regimes whose boundary depends on AI's technology structure (proprietary vs. commodity) and labor market institutions (rent-sharing elasticity, asset concentration). A scenario analysis via Method of Simulated Moments, matching six empirical targets, disciplines the model's quantitative magnitudes; a sensitivity decomposition reveals that the five non-Gini moments identify mechanism rates but not the aggregate sign, which at the calibrated parameters is pinned by and , while AI's technology structure ( vs. ) independently crosses the boundary. The contribution is the mechanism -- not a verdict on the sign. Occupation-level regressions using BLS OEWS data (2019--2023) illustrate why such data cannot test the model's task-level predictions. The predictions are testable with within-occupation, within-task panel data that do not yet exist at scale.
Paper Structure (50 sections, 7 theorems, 15 equations, 2 figures, 8 tables)

This paper contains 50 sections, 7 theorems, 15 equations, 2 figures, 8 tables.

Table of Contents

  1. The Paradox
  2. What this paper contributes.
  3. What the model says.
  4. Positioning.
  5. Model
  6. How AI Changes Task Production
  7. Skill Homogenization Is Arithmetic, but the Economics Is Not
  8. Contrary evidence, scope, and the multiplicative case.
  9. Why Education Returns Decline Unevenly
  10. Credential Inflation as a Screening Response: A Prediction
  11. Information structure.
  12. The Concentrating Channel: Disaggregating Complementary Assets
  13. Why the shift from $\eta_0$ to $\eta_1$ is AI-specific.
  14. Distinguishing from standard capital deepening.
  15. Worker-Firm Assignment
  16. ...and 35 more sections

Key Result

Proposition 1

For any AI-augmentable task $z \in \mathcal{S}(A_t)$, the coefficient of variation (CV) of task output across workers is strictly decreasing in AI capability: where $\mu_h = \mathbb{E}[h_i]$ and $\sigma_h = \sqrt{\text{Var}(h_i)}$. The proportional gain from AI, $\alpha(z)A_t / (h_i\phi(z))$, is strictly decreasing in $h_i$.

Figures (2)

  • Figure 1: Model-implied mechanism chain from AI capability to inequality. The equalizing channel operates through within-task compression. The concentrating channel operates through rising returns to complementary assets. Net inequality depends on AI's technology structure (proprietary vs. commodity) and labor market institutions ($\xi$, $\mathrm{Gini}(K)$).
  • Figure 2: Net inequality effect in $(\eta_1, \mathrm{Gini}(K_0))$ space. Blue: equalizing channel dominates. Red: concentrating channel dominates. Solid black: knife-edge boundary at baseline. Dashed/dotted: boundaries under high/low AI scenarios. Star: calibrated value.

Theorems & Definitions (14)

  • Proposition 1: Skill homogenization
  • proof
  • Proposition 2: Homogenization survives partial complementarity
  • Proposition 3: Bifurcating education returns
  • proof
  • Remark 1: Additive vs. multiplicative human capital
  • Definition 1: Diagnostic variance
  • Proposition 4: Credential inflation from screening degradation
  • proof
  • Lemma 1: Two-sector microfoundation for $\eta'(A) > 0$
  • ...and 4 more