Do People Trust ChatGPT to Write Accurate Statements?

Last registered on July 23, 2024

Pre-Trial

Trial Information

General Information

Title
Do People Trust ChatGPT to Write Accurate Statements?
RCT ID
AEARCTR-0011546
Initial registration date
June 12, 2023

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
June 23, 2023, 4:20 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
July 23, 2024, 3:57 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Region

Primary Investigator

Affiliation
Samford University

Other Primary Investigator(s)

PI Affiliation
George Mason University

Additional Trial Information

Status
Completed
Start date
2023-06-08
End date
2023-08-31
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
In this research paper, we aim to investigate the extent to which humans trust the output of LLMs to be factually accurate. While LLMs have showcased impressive capabilities in understanding and generating text, concerns have been raised regarding the potential for misinformation, biases, or inaccuracies in their responses. To explore this issue, we ask human subjects to rate the accuracy of statements written by humans and ChatGPT. We vary whether each statement was written by a human or by ChatGPT, and we also vary whether we inform participants about who/what wrote the statements.
External Link(s)

Registration Citation

Citation
Buchanan, Joy and William Hickman. 2024. "Do People Trust ChatGPT to Write Accurate Statements?." AEA RCT Registry. July 23. https://doi.org/10.1257/rct.11546-1.2
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2023-06-08
Intervention End Date
2023-08-31

Primary Outcomes

Primary Outcomes (end points)
The dependent variable is a choice by subjects to Trust, Fact-Check, or Not Trust
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We ask human subjects to rate the accuracy of statements written by humans and ChatGPT.
Experimental Design Details
We vary whether each statement was written by a human or by ChatGPT, and we also vary whether we inform participants about who/what wrote the statements.
Randomization Method
Each survey participant is randomly started into one of the information treatments by the computer.
Randomization Unit
Individual
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
450
Sample size: planned number of observations
450
Sample size (or number of clusters) by treatment arms
150 will see AI-informed. 150 will see human-informed. There will be about 75 participants in each uninformed treatment.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
Samford University Institutional Review Board
IRB Approval Date
2023-03-28
IRB Approval Number
EXMT-B-23-S-2

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Yes
Data Collection Completion Date
August 01, 2023, 12:00 +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
Final Sample Size (or Number of Clusters) by Treatment Arms
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Abstract
We explore whether people trust the accuracy of statements produced by large language models (LLMs) versus those written by humans. While LLMs have showcased impressive capabilities in generating text, concerns have been raised regarding the potential for misinformation, bias, or false responses. In this experiment, participants rate the accuracy of statements under different information conditions. Participants who are not explicitly informed of authorship tend to trust statements they believe are human-written more than those attributed to ChatGPT. However, when informed about authorship, participants show equal skepticism towards both human and AI writers. Informed participants are, overall, more likely to choose costly fact-checking. These outcomes suggest that trust in AI-generated content is context-dependent.
Citation
Buchanan, Joy and William Hickman (2024) "Do people trust humans more than ChatGPT?" Journal of Behavioral and Experimental Economics, Volume 112, 102239.

Reports & Other Materials