Assessing the Potential of Generative Agents in Crowdsourced Fact-Checking

Luigia Costabile a , Gian Marco Orlando a , Valerio La Gatta b , Vincenzo Moscato a · 2025

Read Assessing the Potential of Generative Agents in Crowdsourced Fact-Checking by Luigia Costabile a , Gian Marco Orlando a , Valerio La Gatta b , Vincenzo Moscato a (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-19940.txt.

Description

Luigia Costabile and co-authors test a 2025 idea: can AI agents that mimic people help crowds check facts online? Crowd checking is cheap and now used by big platforms, but crowd calls vary in quality and bias. The team built fake crowds of agents with mixed ages, jobs, and views.

Each agent pulled proof, scored claims on set quality points, and gave a final truth call. Agent crowds beat human crowds on truth calls, agreed with each other more, and bent less to social pressure. The results hint at a real role for such agents beside human checkers.

Catalog record

CREATORLuigia Costabile a , Gian Marco Orlando a , Valerio La Gatta b , Vincenzo Moscato a
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE89 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-19940

Browse the full library · About the Tower of Babel