Scaling Laws for Economic Productivity: Experimental Evidence in White-Collar Professional Tasks

How to cite

merali, Ali. 2025. "Scaling Laws for Economic Productivity: Experimental Evidence in White-Collar Professional Tasks." AEA RCT Registry. AEARCTR-0015945. https://doi.org/10.1257/rct.15945-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Ali merali*

* Primary investigator

Registry ID
AEARCTR-0015945
DOI
https://doi.org/10.1257/rct.15945-1.0
First registered
Last updated
Status
In development
Trial dates
2025-05-12 – 2025-07-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/15945

Abstract

This study investigates the productivity impacts of sixteen generative artificial intelligence (AI) technologies with varying training compute sizes. By analyzing scaling laws, we measure how economic impacts evolve as the size of generative AI models increases. Our research is conducted in a real-world context, employing professional consultants, managers, data analysts, solicitors, and software engineers to perform tasks akin to their routine professional work. Additionally, we explore the relationship between the skill-bias of technological change and the degree of automation. Our findings offer critical insights into the scaling effects of AI on productivity and the dynamics of skill-biased technological advancements in the labor market.

Intervention

See Analysis Plan for full regression specifications/ definitions:
1. log Productivity by log model compute (Main)
2. Interaction between respondent baseline ability and log model compute (Main 2)

Experimental design

Recruiting 200 professional across five different professions (managers, consultants, data analysts, solicitors, and software engineers. Each professional will first complete a baseline task of ability. Then they will complete three short follow-up tasks. In each follow-up task individuals will equally be split into one of sixteen groups with different generative AI models. They will also all complete some short questions (eg. on their experience as a professional, familiarity with AI tools, etc.)

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
Individual
Clusters
250 (50 individuals each across 5 different professions)
Observations
750= 250 professionals completing three tasks each
Arms
750 observations/ 16 treatments= roughly 47 tasks per treatment arm
Primary outcomes
log productivity (time taken) earnings per hour grade These three interacted with a dummy variable on ability

Cited by

No citation data available for this registration yet.

Keywords