VeriFight: Fighting for the truth
A browser plugin for fake news identification
With the use of AI and big data analytics, Verifight provides a bundle of signals for fake news detection, developed as a privacy-first browser plugin for everyday secure usage.
Team membersDemetris Paschalides, Chrysovalantis Christodoulou
Members roles and background
Chrysovalantis Christodoulou received his BSc (2019) in the Department of Computer Science at the University of Cyprus. Currently, he is a Master's Student also in the Computer Science department at the University of Cyprus. He also works as a Researcher at the Laboratory of Internet Computing. During his studies, he participated in two competitions for start-up ideas and won both of them with his team. His research interests are related to Fake News Detection, Natural Language Processing, Feature Selection, and Machine learning.
Demetris Paschalides is a researcher and a Ph.D. candidate at the Laboratory of Internet Computing at the University of Cyprus. His work mainly focuses on the behavioral analysis and information extraction of users in online social networks with data analysis, natural language processing, and machine learning techniques. Demetris has worked on subjects like online hate speech and fake news detection, with multiple additional side-projects of data analysis and web development.
VeriFight is developed as a browser plugin, to conveniently work in the background, protecting online readers from disinformation, rumors, and hoaxes. VeriFight uses lightweight and secures API to analyze the article actively being read by the user and timely warn him/her of any suspicion on misinformation. As we take the user's privacy seriously, no data are being withheld on Verifight's server for personalization, analysis, or whatsoever. From a more technical perspective, VeriFight combines a series of signals to produce a suspicion score around the article that is being analyzed. These signals correspond to actual steps that a person with proper media literacy would apply but in an automated manner. First, the origin of the article is evaluated. Based on a series of known news media domains that produce false information, and promote conspiracy theories, VeriFight identifies if the current domain exists within those, and generates information on the factuality of the domain. Furthermore, different linguistic characteristics are analyzed from the content and title of the news article. Is the title a click-bait? What is the intention of the article? To scare? To frighten? Is the article bias? These are a sample of questions that are being answered. Last but not least, VeriFight identifies any ongoing rumors and conspiracy theories regarding COVID-19 within the articles. These rumors are continuously updated from several databases around the world, based on the worldwide initiative against the COVID 'infodemic'. All of these are produced from AI models that look at the article, and the use of Big Data Analytics, but are also steps any person should take in validating the veracity of an article.
In recent years, the fake news crisis has worsened and impacted our daily lives; from politics, enterprises to human lives, including the COVID-19 pandemic and series of conspiracy theories and fake cures. Despite any clear pattern of misinformation, people still fall victim to such news, indicating the lack of media literacy and critical thinking. We tackle this, by mimicking the media literate approach for identifying fake news in an automated way.
Solution target group
We have two separated target groups:
1. The first one represents daily user which are falling victims to misinformation
2. The second one refers to News agencies / Governments/ Other Organizations who would like to verify their information using our extensive set of big data analytics.
VeriFight aims to alter how people consume information, especially misinformation and fake news. One of our impact factors is the number of downloads and usage analytics of the Verifight browser plugin. The most important impact factor is based on click analysis and crowdsourcing, indicating how well Verifight can manage to prevent the spread and penetration of fake news.
Solution tweet textStrap in to the battle against fake news #no#more#fakenews#verifightPlugin.
There is a variety of browser plugins trying to identify fake news and warn users. However, most of the approaches ignore users' privacy and they are asking for accounts and access to users' browser history. The first priority of our approach is to preserve this privacy and therefore we will not keep anything about the user and what pages have visited. The second disadvantage that most of our competitors have is the lack of different signal combinations. Leveraging a blacklist with fake domains or focusing only on Tweets does not meet the need for real mitigation of misinformation. Our extension combines a variety of signals coming from big data analytics and machine learning in order to advise the user about the veracity of an article. Considering both of the previously mentioned advantages our solution is innovative and aims to disrupt the current approaches for Fake News Detection.
Currently, Verifight is tailored in COVID-19 specific fake news identification but can be transition easily to identify fake news in general. Furthermore, Verifight's infrastructure and big data processing, along with the series of machine learning models, can be transferred to enterprises and the fraud and risk management market, as well as market and brand analytics, focusing on specific topics.
As we mentioned before we have two different groups of users. The first group which includes daily users will not provide us much profit, because we are planning to provide the tool for free. However, can be used as a means of advertisement for our product and the possibilities. Also, the crowdsourcing input is crucial, for feedback to our machine learning models, and impact factor calculations. After we make a name for ourselves, the enterprise edition tier-based sustainability plan is taken into place. Enterprises, organizations, and even governments will pay a subscription fee to use Verifight infrastructure and an extensive set of tools for analytic insights on fake news and misinformation.
Solution team work
We have been working as a team for quite some time now, and have flourished during the DigiEduHack event. Our technical and academic knowledge has provided us with the tools and proper collaboration to achieve, to our knowledge, great results in such a short period of time. From the experience we have so far, greatness awaits in the future of our team.
* Climate-KIC publishes the proposed solutions developed during the DigiEduHack event solely for the purposes of facilitating public access to the information concerning ideas and shall not be liable regarding any intellectual property or other rights that might be claimed to pertain to the implementation or use any of the proposed solutions shared on its website neither does it represent that it has made any effort to identify any such rights. Climate-KIC cannot guarantee that the text of the proposed solution is an exact reproduction of the proposed solution. This database is general in character and where you want to use and develop a proposed solution further, this is permitted provided that you acknowledge the source and the team which worked on the solution by using the team’s name indicated on the website.