Affiliation:
1. Michigan State University, East Lansing, MI, USA
2. Stanford University, Stanford, CA, USA
Abstract
To the human eye, AI-generated outputs of large language models have increasingly become indistinguishable from human-generated outputs. Therefore, to determine the linguistic properties that separate AI-generated text from human-generated text, we used a state-of-the-art chatbot, ChatGPT, and compared how it wrote hotel reviews to human-generated counterparts across content (emotion), style (analytic writing, adjectives), and structural features (readability). Results suggested AI-generated text had a more analytic style and was more affective, more descriptive, and less readable than human-generated text. Classification accuracies of AI-generated versus human-generated texts were over 80%, far exceeding chance (∼50%). Here, we argue AI-generated text is inherently false when communicating about personal experiences that are typical of humans and differs from intentionally false human-generated text at the language level. Implications for AI-mediated communication and deception research are discussed.
Subject
Linguistics and Language,Sociology and Political Science,Anthropology,Language and Linguistics,Education,Social Psychology
Reference61 articles.
1. Agrawal A., Gans J., Goldfarb A. (2017). How AI will change the way we make decisions. Harvard Business Review. https://hbr.org/2017/07/how-ai-will-change-the-way-we-make-decisions.
2. Fitting Linear Mixed-Effects Models Usinglme4
3. Benoit K., Watanabe K., Wang H., Lua J. W., Kuha J. (2021). quanteda.textstats: Textual statistics for the quantitative analysis of textual data (0.95) [Computer software]. https://CRAN.R-project.org/package=quanteda.textstats.
4. Discourse on the Move
Cited by
4 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献