The state of human-centered NLP technology for fact-checking

Misinformation threatens modern society by promoting distrust in science, changing narratives public health, heightening social polarization, and disrupting democratic elections financial markets, among a myriad of other societal harms. To address this, growing cadre professional fact-checkers journalists provide high-quality investigations into purported facts. However, these largely manual efforts have struggled to match the enormous scale problem. In response, body Natural Language Processing (NLP) technologies been proposed for more scalable fact-checking. Despite tremendous growth such research, however, practical adoption NLP fact-checking still remains its infancy today. this work, we review capabilities limitations current Our particular focus is further chart design space how can be harnessed refined order better meet needs human fact-checkers. do so, key aspects NLP-based fact-checking: task formulation, dataset construction, modeling, human-centered strategies, as explainable models human-in-the-loop approaches. Next, efficacy applying tools assist We recommend that future research include collaboration with fact-checker stakeholders early on well incorporation practices model development, guide technology development use adoption. Finally, advocate benchmark supporting extrinsic evaluation technologies.