Can AI help teachers? Experimental Evidence on the effects of Generative AI-use and AI training for teachers' work

How to cite

Fervers, Lukas; Appel, Philipp. 2024. "Can AI help teachers? Experimental Evidence on the effects of Generative AI-use and AI training for teachers' work." AEA RCT Registry. AEARCTR-0014952. https://doi.org/10.1257/rct.14952-1.0

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

Lukas Fervers* , Philipp Appel

* Primary investigator

Registry ID
AEARCTR-0014952
DOI
https://doi.org/10.1257/rct.14952-1.0
First registered
Last updated
Status
In development
Trial dates
2024-12-05 – 2025-01-31
Countries
Germany ()
JEL codes
I20; I2; C90;
Registry record
http://www.socialscienceregistry.org/trials/14952

Abstract

Artificial intelligence, in particular Large Language Models (LLMs), are commonly considered being a path-breaking innovation. While possible applications and rising research questions are manifold, one straightforward question of very general interest is whether LLMs can support humans in fulfilling job-related tasks, hereby improving the quality and quantity of the output. In this study, we will focus on the impact of LLM and LLM training on teachers in adult education. To this end, we design an experimental study in which we ask participants to complete four tasks typical for teaching preparation and communication with course participants. We vary both access to an assisting LLM as well as being trained in the use of LLMs before completing the tasks. This will provide insights about whether and to what extend adult education teachers can profit from the use of LLMs or being trained in their use to improve the quality of teaching material and/or increase the efficiency in teaching preparation.

Intervention

Study participants receive an introduction into the use of LLMs. While some components of the introduction are rather general, it is particularly tailored for the preparation of teaching material in adult education. The training is thus supposed to support participants in using LLMs to enhance the quality of teaching material and reduce the time needed for its preparation.

Experimental design

The experiment is conducted as online experiment with delayed control group design.

Design details

Randomization method
Randomization is conducted by the software labvanced when participants enter the intervention (simple randomization)
Randomization unit
Individuals
Clusters
No clustering, invitation and randomization is conducted at the individual level
Observations
We invite 800 persons to participate in the study. We expect 160 participants in total.
Arms
We expect to have between 50 and 60 persons per treatment arm
Primary outcomes
Overall performance in completing the tasks
Secondary outcomes
Disaggregate analyses for separate tasks Political attitudes

Cited by

No citation data available for this registration yet.

Keywords