Betsy Sinclair | Searcharxiv

arXiv2024

Towards Generalizable AI-Assisted Misinformation Inoculation: Protecting Confidence Against False Election Narratives

We present a generalizable AI-assisted framework for rapidly generating effective "prebunking" interventions against misinformation. Like mRNA vaccine platforms, our approach uses a stable template structure that can be quickly adapted to counter emerging false narratives. In a preregistered two-wave experiment with 4,293 U.S. registered voters, we test this framework against politically-charged election misinformation -- one of the most challenging domains for misinformation intervention. Our design directly tests scalability by comparing human-reviewed and purely AI-generated inoculation messages. We find that LLM-generated prebunking significantly reduced belief in election rumors (persisting for at least one week) and increased confidence in election integrity across partisan lines. Purely AI-generated messages proved as effective as human-reviewed versions, with some achieving larger protective effects, demonstrating that effective misinformation inoculation can be achieved at machine speed without proportional human effort, offering a scalable defense against the accelerating threat of false narratives across all domains.

econ.GN↗

arXiv2008

Competition between Discrete Random Variables, with Applications to Occupancy Problems

Consider $n$ players whose "scores" are independent and identically distributed values $\{X_i\}_{i=1}^n$ from some discrete distribution $F$. We pay special attention to the cases where (i) $F$ is geometric with parameter $p\to0$ and (ii) $F$ is uniform on $\{1,2,...,N\}$; the latter case clearly corresponds to the classical occupancy problem. The quantities of interest to us are, first, the $U$-statistic $W$ which counts the number of "ties" between pairs $i,j$; second, the univariate statistic $Y_r$, which counts the number of strict $r$-way ties between contestants, i.e., episodes of the form ${X_i}_1={X_i}_2=...={X_i}_r$; $X_j\ne {X_i}_1;j\ne i_1,i_2,...,i_r$; and, last but not least, the multivariate vector $Z_{AB}=(Y_A,Y_{A+1},...,Y_B)$. We provide Poisson approximations for the distributions of $W$, $Y_r$ and $Z_{AB}$ under some general conditions. New results on the joint distribution of cell counts in the occupancy problem are derived as a corollary.

math.PR↗