AMITT (Adversarial Misinformation and Influence Tactics and Techniques) framework for describing disinformation incidents. AMITT is part of misinfosec - work on adapting information security practices to help track and counter misinformation - and is designed as far as possible to fit existing infosec practices and tools.
Label each twitter account with label + confidence score for primary:
-- long/lat <-- just start here...
-- country
-- if US: state, zipcode, city/county (?)
Helpful info:
-- 1% of tweets have geo data
-- sample use case: "who was early/late to adopting Masks4All? who is currently resisting / not activating?"
Unclear: People move and have separate home/work... so how does recency play
A minimum-dependency ECMAScript client library and CLI tool for Parler – a "free speech" social network that accepts real money to buy "influence" points to boost organic non-advertising content
A data set regarding news veracity comprised of data from disparate sources; including Facebook, several online news outlets, and Disqus. Published at ICWSM-18.
Supplemental materials for Karduni et al. (ICWSM 2018) - "Can You Verifi This? Studying Uncertainty and Decision-Making about Misinformation in Visual Analytics"
This is the repository for the paper "The Rise of Guardians: Fact-checking URL Recommendation to Combat Fake News" accepted as a full regular paper by SIGIR 2018
This repository provides several Python scripts to extract tweets possibly containing health misinformation. These tweets have replies semantically similar with official advice from health authorities such as the WHO and the CDC, inspired by the observations that some tweets were corrected by volunteer fact checkers with accurate information.
Health misinformation can spread rapidly during a crisis, as in the case of the COVID-19 pandemic. We provide a tool that collects Twitter posts that likely contain misinfo for use by fact checkers.
We believe that curbing this subset of fake news would involve providing the necessary tools to journalists and users who take the time to sift through trustworthy content. In this light, our solution is to devise a methodology and program to assist a journalist to identify if an image and associated title are trustworthy. The fundamentals of our approach are based on corroboration of news, i.e, the image must be used in credible news websites with text similar to that of the associated title for the image and text to be reliable. If not, we recommend that the journalist (or user) perform a human verification procedure.
Label each twitter account with label + confidence score for primary:
-- long/lat <-- just start here...
-- country
-- if US: state, zipcode, city/county (?)
Helpful info:
-- 1% of tweets have geo data
-- sample use case: "who was early/late to adopting Masks4All? who is currently resisting / not activating?"
Unclear: People move and have separate home/work... so how does recency play