THE ASYMMETRIC IMPACT OF GENERATIVE ARTIFICIAL INTELLIGENCE ON THE RETURNS TO COGNITIVE LABOR: A QUASI-EXPERIMENTAL STUDY OF HIGH-SKILLED FREELANCE MARKETS

Anika Singh · International Journal of Economics · 22 Jun 2026

Where this sits in the debate

Will AI destroy more jobs than it creates? · its finding: it depends (why studies differ)
Will AI widen inequality? · its finding supports the claim (why studies differ)

Direction of the finding as read from the abstract by an AI model (AI-written).

Key takeaway AI-written

Using a difference-in-differences design on global freelance platforms, this study finds generative AI compresses the productivity gap between novice and expert cognitive workers—lowering the skill premium and entry-level wages in coding, technical writing, and design—while raising the value of human-in-the-loop supervision and strategic work; it also models the elasticity of substitution between human judgment and algorithmic output. This is direct quasi-experimental evidence for the leveling-vs-displacement question, and its implication that junior-role automation breaks internal promotion ladders is central to debates about how firms will rebuild talent pipelines.

The paper in three steps AI-written

Written by an AI agent from the paper's abstract, which is the primary source and is quoted below.

The question

How does generative AI change the returns to skill among high-skilled freelance workers in coding, technical writing and design?

The approach

A difference-in-differences design on freelance platforms worldwide, plus a simulation of the elasticity of substitution between human work and algorithm output.

The findings

Early evidence indicates AI narrows the productivity gap between beginner and expert workers, which could lower wages for entry-level cognitive jobs and raise the value of human supervision and strategic planning.

How to read it: observational. Analysis of real-world firm, worker, or market data — realistic settings; causal claims depend on the identification strategy.

Limitations and counterpoints AI-written

The abstract gives no sample size, platforms, period or effect sizes, and calls the evidence early.

Abstract

The study explores how the labor market would be upset by the widespread use of Large Language Models (LLMs) and multimodal AI in the professional creative and technical industries.Whereas past instances of automation involved mainly routine manual activity, the present wave aims at non-routine cognitive activity, in this case, coding, technical writing, and graphic design.This paper measures the change in the skill premium using a difference-indifferences design on freelance websites worldwide.Early evidence indicates an offsetting effect in that AI is reducing the productivity difference between beginner and expert employees, which has the potential to lower wages on entry-level thinking jobs and soar the worth of the so-called human-in-theloop supervision and strategic planning.We simulate the Elasticity of Substitution between human intuition and the output of an algorithm, which offers a template of how firms ought to reorganize internal promotion ladders when the positions of the socalled juniors are significantly automated.

About this publication

Academic journal. Published in an academic journal — typically peer-reviewed before publication, though indexing alone doesn't guarantee review depth.

Published: 2026 (4 months ago)
Citations: 0 (OpenAlex)