Kauwa Kaate Fact Checker from IIT Bombay

Developer's DescriptionBy IITB InfolabThe goals of the Kauwa Kaate project at IIT Bombay are broadly to combat the fake news epidemic that has affected the world today.The goals of the Kauwa Kaate project at IIT Bombay are broadly to combat the fake news epidemic that has affected the world today. Much of the fake news circulation in India happens via WhatsApp, and a key goal of the project is to make it easy for users to fact check forwards that they receive.We have developed a fact checker that you can access using the Kauwa Kaate Fact Check App on PlayStore, or on the web at http://kkfactcheck.in or https://www.cse.iitb.ac.in/infolab/fakenews/factcheck.You can copy-paste text from WhatsApp to our app, or share images/videos from WhatsApp to our app (or website).Our fact checker checks your text (articles), images, and videos, against1. A predefined set of India focused fact checking sites such as https://altnews.in, https://boomlive.in and https://check4spam.com; more such sites are being added. Since we need to match potentially long articles, along with image matching, and these features are not supported by web search engines, we crawl these sites regularly and perform search on our server on the crawled data.2. A predefined set of (relatively trusted) news sites such as Times of India and The Hindu, to see if the article being fact checked has appeared on one of these sites, which can then be labeled as (relatively) trustworthy.Support for image/video search is an important part of our project, which cannot be done as easily using google search.If an article/image/video does not match either of the above cases, there are a few automated checks that can be done, and we are working on these currently. More importantly, we are working on making the data about searches available to fact checkers.Note that like Google, we retain your search queries, but we do not retain any information about who carried out what search, so your search is anonymous.A long term research goal of the project is to do more on this front, but currently it is limited. For example, we wish to check against a set of sites known to be biased on a particular topic (say politics), and tell the user that a forward appears on such a site, and they should dig deeper and not necessarily accept it as is.