EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

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

Matthew R DeVerna iD, Harry Yaojun Yan iD, Kai-Cheng Yang iD, Filippo Menczer iD

Proceedings of the National Academy of Sciences · 10.12.2024 · 121 (50)

DOI: 10.1073/pnas.2322823121 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the National Academy of Sciences, online erschienen am 04.12.2024. Verfasst von Matthew R DeVerna und 3 weiteren Autor:innen. Beteiligte Einrichtung: Observatory on Social Media, Indiana University, Bloomington, IN 47408. Hauptthemen laut MEDLINE: Language. Die Arbeit wurde bisher 54-mal zitiert (Crossref) und nennt 80 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: John S. and James L. Knight Foundation (KF), Volkswagen Foundation (VolkswagenStiftung).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

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.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

LanguageHumansCommunicationFemaleMaleInformation Dissemination

Stichwörter der Autor:innen

AIfact-checkingheadline discernmentlarge language modelsmisinformation

Autor:innen und Institutionen

  • Observatory on Social Media, Indiana University, Bloomington, IN 47408.
  • Stanford Social Media Lab, Stanford University, Stanford, CA 94305.
  • Network Science Institute, Northeastern University, Boston, MA 02115.

ORCID-Profile: Matthew R DeVerna, Harry Yaojun Yan, Kai-Cheng Yang, Filippo Menczer

Förderung

  • John S. and James L. Knight Foundation (KF)
  • Volkswagen Foundation (VolkswagenStiftung)

Zitieren

APA

DeVerna, M. R., Yan, H. Y., Yang, K. C., & Menczer, F. (2024). Fact-checking information from large language models can decrease headline discernment. Proceedings of the National Academy of Sciences, 121(50). https://doi.org/10.1073/pnas.2322823121

Vancouver

DeVerna MR, Yan HY, Yang KC, Menczer F. Fact-checking information from large language models can decrease headline discernment. Proceedings of the National Academy of Sciences. 2024;121(50). doi:10.1073/pnas.2322823121

BibTeX
@article{DeVerna2024,
  author = {DeVerna, Matthew R and Yan, Harry Yaojun and Yang, Kai-Cheng and Menczer, Filippo},
  title = {Fact-checking information from large language models can decrease headline discernment},
  journal = {Proceedings of the National Academy of Sciences},
  year = {2024},
  volume = {121},
  number = {50},
  doi = {10.1073/pnas.2322823121}
}
Kostenlos starten →