Abstract

Can large language models, a form of artificial intelligence (AI), generate persuasive propaganda? We conducted a preregistered survey experiment of US respondents to investigate the persuasiveness of news articles written by foreign propagandists compared to content generated by GPT-3 davinci (a large language model). We found that GPT-3 can create highly persuasive text as measured by participants’ agreement with propaganda theses. We further investigated whether a person fluent in English could improve propaganda persuasiveness. Editing the prompt fed to GPT-3 and/or curating GPT-3’s output made GPT-3 even more persuasive, and, under certain conditions, as persuasive as the original propaganda. Our findings suggest that propagandists could use AI to create convincing content with limited effort.

Details

Title
How persuasive is AI-generated propaganda?
Author
Goldstein, Josh A 1   VIAFID ORCID Logo  ; Chao, Jason 2   VIAFID ORCID Logo  ; Grossman, Shelby 2   VIAFID ORCID Logo  ; Stamos, Alex 2 ; Tomz, Michael 3   VIAFID ORCID Logo 

1  Center for Security and Emerging Technology, Georgetown University , Washington, DC 20001 , USA 
2  Stanford Internet Observatory, Stanford University , Stanford, CA 94305 , USA 
3  Department of Political Science and Stanford Institute for Economic Policy Research, Stanford University , Stanford, CA 94305 , USA 
Publication year
2024
Publication date
Feb 2024
Publisher
Oxford University Press
e-ISSN
27526542
Source type
Scholarly Journal
Language of publication
English
DOI
https://doi.org/10.1093/pnasnexus/pgae034
ProQuest document ID
3191895054
Copyright
© The Author(s) 2024. Published by Oxford University Press on behalf of National Academy of Sciences. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.