Reproduce “When Imperfect Automated Fact-Checking Backfires: Social Learning in the Presence of Misinformation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Zixuan Liu, Ming Cao
The authors model sequential judgments with private, free, imperfect automated checks and costly, accurate manual checks; successors observe actions, not verdicts. They define welfare as the probability of a correct cascade, when later users follow accumulated actions. Under theoretical conditions, adding automation lowers welfare for intermediate-reliability articles. In one simulation setup, average welfare crosses manual-only near 0.81 balanced accuracy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.