Does AI Deepen the Divide? Examining the Effects of AI-Generated Content on Confirmation Bias and Opinion Polarization

How to cite

LYU, Aoqing; Zhang, Xu. 2026. "Does AI Deepen the Divide? Examining the Effects of AI-Generated Content on Confirmation Bias and Opinion Polarization." AEA RCT Registry. AEARCTR-0018421. https://doi.org/10.1257/rct.18421-1.0

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

Aoqing LYU* , Xu Zhang

* Primary investigator

Registry ID
AEARCTR-0018421
DOI
https://doi.org/10.1257/rct.18421-1.0
First registered
Last updated
Status
In development
Trial dates
2026-04-29 – 2026-05-07
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18421

Abstract

This study examines whether AI-generated content affects confirmation bias in information processing. Using an online experiment, it compares responses to AI-generated and human-written content across controversial social issues. The analysis focuses on two dimensions of confirmation bias, namely information interpretation and information selection, and considers whether source-related features moderate these effects. The study aims to provide evidence on whether AI-generated content mitigates or amplifies confirmation bias and, more broadly, its potential role in shaping opinion polarization.

Intervention

The intervention is an incentivized online information-exposure experiment on controversial social issues. Participants are exposed to both AI-generated and human-written content, as well as to supportive and opposing viewpoints, with the order of the six topics randomized. The study uses a Bayesian Truth Serum (BTS) incentive scheme to encourage truthful reporting.

Experimental design

This study is an incentivized online randomized experiment examining how AI-generated versus human-written information affects confirmation bias and opinion polarization across controversial social issues. The design combines within-subject and between-subject variation. Participants are exposed within-subjects to content that varies by source and stance, while source-related presentation features are assigned between subjects and balanced across treatment conditions. Truthful reporting is incentivized using Bayesian Truth Serum (BTS).

The experiment consists of two main components. In the first, participants are exposed to multiple articles that vary in source and viewpoint, with randomized presentation order. In the second, participants make choices among titles that vary along similar dimensions and are later exposed to related content. The design additionally varies whether source-related features remain aligned across stages of exposure. Topic order is randomized, and treatment assignments are balanced across participants.

Design details

Randomization method
Randomization was implemented by computer. Topic order and balanced treatment assignments were pre-generated in Excel using random numbers. Participants were then automatically assigned by Credamo to one of the experimental conditions, with all subsequent randomized elements determined by the pre-generated assignment.
Randomization unit
individual
Clusters
512 individual participants
Observations
512 individual participants
Arms
128 participants in labeled-consistent, 128 in labeled-inconsistent, 128 in unlabeled-consistent, and 128 in unlabeled-inconsistent conditions. Within each treatment arm, participants are exposed to both AI-generated and human-written content and to both supportive and opposing stances.
Primary outcomes
The primary outcome variables are: (1) title selection, measured by which title is chosen from the four available options for each topic (AI-supportive, AI-opposing, human-supportive, human-opposing); and (2) post-exposure evaluations and attitudes following article reading, title selection, and the final related-article stage.

Cited by

No citation data available for this registration yet.

Keywords