Impact of the introduction of AI technology on the employment environment: International survey

How to cite

takahashi, koji; Aoki, Yusuke; Takada, Yuya. 2024. "Impact of the introduction of AI technology on the employment environment: International survey." AEA RCT Registry. AEARCTR-0013927. https://doi.org/10.1257/rct.13927-1.0

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

koji takahashi* , Yusuke Aoki , Yuya Takada

* Primary investigator

Registry ID
AEARCTR-0013927
DOI
https://doi.org/10.1257/rct.13927-1.0
First registered
Last updated
Status
Completed
Trial dates
2024-05-20 – 2024-06-03
Countries
Japan (), United States of America ()
JEL codes
E24, E31, O30
Registry record
http://www.socialscienceregistry.org/trials/13927

Abstract

We investigate how perceptions of generative AI's impact on the labor market affect macroeconomic expectations and individual behavioral intentions. Participants are shown different predictions about AI job replacement and are then asked about their expectations for CPI inflation, GDP growth, and their willingness to engage with AI in their work and studies.

Intervention

We investigate how perceptions of the impact of generative AI on the labor market influence macroeconomic expectations and wage change formation. First, participants are asked about the extent to which generative AI will replace workers in the future. Then, participants are randomly assigned to one of three groups. Group A is shown a prediction by an expert stating that AI will replace 14% of all jobs. Group B is shown a prediction that AI will replace a higher percentage, 47% of all jobs. Group C, serving as a placebo, is shown a future prediction about an unrelated astronomical phenomenon (Halley's Comet). For these three groups, we will examine changes in macroeconomic expectations by asking participants to predict CPI inflation and GDP growth rates before and after the treatment. Additionally, we will ask respondents about their willingness to study AI in the future and the extent to which they want to actively use AI in their work, to observe how individual behavioral intentions toward AI are affected.

Experimental design

The surveys were conducted via the internet, targeting large-scale survey monitors.
Respondents were individuals who had previously registered with the survey company and could win cash or gift vouchers for completing the survey. The sample was equally distributed by sex, age, and job category (16 groups). Sex was divided into two groups: female and male (2 groups). Age was divided into two groups: 20 to 39 years old and 40 to 59 years old (2 groups). Job categories included Sales and Administrative, Engineering, Planning and Specialist, and Creative (4 groups). Only full-time company employees were invited.
We randomly assigned respondents into three groups to evaluate the impact of generative AI on job replacement. Group A, labeled as the "Small impact of AIs" group, received the following treatment: “A well-known study on AI estimates that 14% of current jobs could be replaced by generative AI in the future.” Group B, labeled as the "Large impact of AIs" group, received this treatment: “A well-known study on AI estimates that 47% of current jobs could be replaced by generative AI in the future.” Group C served as the placebo group and was informed with unrelated information as a control: “According to astronomical studies, the next time Halley's Comet will approach the Earth will be in July 2061.”

Respondents were then asked the questions about AI and macroeconomy both before and after receiving the treatment. This allowed us to measure any changes in their responses due to the information provided. This presentation focuses on the findings from Groups A and B.

Design details

Randomization method
randomization done in office by a computer
Randomization unit
individual
Clusters
No cluster
Observations
8000
Arms
1400 treatment with high replacement ratio, 1400 treatment with low replacement ratio, 1400 placebo with being provided with unrelated information
Primary outcomes
expectations of real GDP growth, inflation rate, private investment growth, wages, intention to learn AI, intention to use AI

Cited by

No citation data available for this registration yet.

Keywords