arrow
返回

Fact-checking information from large language models can decrease headline discernment

delete2024-12-04
delete0
delete
OA
AI
M
Matthew DeVerna *
H
Harry Yaojun Yan
K
Kai‐Cheng Yang
F
Filippo Menczer
DOI:10.1073/pnas.2322823121delete
delete原文链接
delete分享
delete收藏
查看原文
摘要

摘要

En 中文
Fact checking can be an effective strategy against misinformation, but its implementation at scale is impeded by the overwhelming volume of information online. Recent AI language models have shown impressive ability in fact-checking tasks, but how humans interact with fact-checking information provided by these models is unclear. Here, we investigate the impact of fact-checking information generated by a popular large language model (LLM) on belief in, and sharing intent of, political news headlines in a preregistered randomized control experiment. Although the LLM accurately identifies most false headlines (90%), we find that this information does not significantly improve participants' ability to discern headline accuracy or share accurate news. In contrast, viewing human-generated fact checks enhances discernment in both cases. Subsequent analysis reveals that the AI fact-checker is harmful in specific cases: It decreases beliefs in true headlines that it mislabels as false and increases beliefs in false headlines that it is unsure about. On the positive side, AI fact-checking information increases the sharing intent for correctly labeled true headlines. When participants are given the option to view LLM fact checks and choose to do so, they are significantly more likely to share both true and false news but only more likely to believe false headlines. Our findings highlight an important source of potential harm stemming from AI applications and underscore the critical need for policies to prevent or mitigate such unintended consequences.
Keyword:
fact-checking
large language models
misinformation
headline discernment
AI
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

P
Proceedings of the National Academy of Sciences of the United States of America
IF:
9.1
论文数:
10.8W
被引数:
96.0W

机构

I
indiana university system
学者数:
4.0W
论文数: 3.5W
被引数: 4.6W
引用论文

引用论文

Changes in Disparity in County-Level Diagnosed Diabetes Prevalence and Incidence in the United States, between 2004 and 2012
err2016-08-03
err0
errOAAI
errSundar S. Shrestha; Theodore J. Thompson; Karen A. Kirtland; Edward W. Gregg; Gloria L. Beckles; Elizabeth T. Luman; Lawrence E. Barker; Linda S. Geiss
err分享
err收藏
err分享
err收藏
Social media behavior is associated with vaccine hesitancy社交媒体行为与疫苗犹豫有关
err2022-09-30
err46
errOAAI
errRathje, Steve; He, James K.; Roozenbeek, Jon; Van Bavel, Jay J.; van der Linden, Sander
err分享
err收藏
Addressing the harms of AI-generated inauthentic content
err2023-07-12
err10
PREAI
errMenczer, Filippo; Crandall, David; Ahn, Yong-Yeol; Kapadia, Apu
err分享
err收藏
Online misinformation is linked to early COVID-19 vaccination hesitancy and refusal
err2022-04-26
err136
errOAAI
errPierri, Francesco; Perry, Brea L.; DeVerna, Matthew R.; Yang, Kai-Cheng; Flammini, Alessandro; Menczer, Filippo; Bryden, John
err分享
err收藏
err分享
err收藏
学者 查看更多内容