NewsBuddy

 Introduction

Fake news has existed ever since the dawn of the printing press, but its impact has increased tremendously in the internet age. It is often used to spread hate and defame people or organisations. Uncredible news articles mislead people and create social conflicts. We have seen political parties use propaganda techniques to manipulate people. Fake news can also be used for disrupting the economy. What makes fake news a challenging problem to solve is identifying, tracking, and controlling uncredible content. 

Problem Statement

The large volume of the news articles posted online makes it difficult to moderate the credibility of the content since professional fact-checking is tough to scale. This is a huge bottleneck in the campaign against misinformation.

Our Journey

In order to come up with an efficient solution for the above problem statement, we followed a timeline. At every step of the timeline, we had a lot of brainstorming sessions and received valuable feedback from Prof. Ponnurangam Kumaruguru and the TAs of the PSOSM course.

User Survey & Interviews

We spent a considerable amount of time conducting user surveys and interviews to accurately understand the needs of the users. We tried to incorporate people from different age groups and backgrounds in our survey so that our solution can cater to the general public better. The users were questioned about how the problem of fake news affects them and what features can help them in deciding the credibility of a news article better.

Survey Analysis

After conducting surveys and interviews, we obtained a lot of insightful data. We used the following design research techniques to organise our insights and ideas.

Competitive Analysis

To find the potential of our application and the current market competition, we did a competitive analysis and formulated our findings in the following table.

Affinity Mapping

We collected all the interview data together and put them in an Affinity Map, which helped us to find the needs of our users.

Personas

Using all of the data collected above, we divided our target audience into 3 main personas to better understand all their needs.

Designing Architecture & Interface

After doing a proper analysis of user survey data, we were able to finalise the overall design and features of our solution. 

• We developed a web application using Django for the backend and HTML, CSS, JS for the frontend.
• A user can log in to the portal and enter the link for any news article. The content from a news article link is extracted using the trafilatura library and rendered as an HTML page.
• The portal allows the user to dispute the parts of the article which he/she feels are misleading. It is necessary for a user to provide a reference for his/her claim.
• The arguments made by other users on the article are visualised using charts to help user judge the article better.
• The Django server manages all the user logins and fetches the comments and annotations.  
• As an incentive for the users, we introduced gamification features like badges and points which a user can gain by annotating articles.
• The overall UI is kept minimalistic for a good reading experience.

With the crowdsourcing approach, everyone can contribute to verify the credibility of content in the news articles. Users can be more confident about the claims of an article when they have been scrutinised by other common people. The platform can also promote community building. In the larger term, a major bottleneck in the spread of misinformation would be removed.

Prototyping & User Evaluation

We designed a prototype of our system and conducted user evaluations to better understand our pitfalls.

To test the usability of our product, we conducted guerilla evaluations with our users and these were some of the feedback that we got.

The requirement of a reference link is a good step to prevent spamming.
The annotations were a little annoying, I wish I could turn them off when I just want to read the article.
The visualisation of public opinions as charts is appreciable.
It would be great if more options are available to customise user profiles.
I like how the platform provides a way to backtrack claims people make on news articles.

Final Product

NewsBuddy is a web application that tackles the problem of fake news using a crowdsourcing approach. It empowers the users to debate upon the correctness of news articles and make an informed decision about an online news article's credibility. Users can agree/dispute with specific claims of an article and can submit reviews backed by certain reliable references (top news channels, reputed sources of information). The disputed sections of an article are highlighted and arguments provided by other users for the same are shown. Gamification features like badges encourage user participation and thereby make the system better.

Demo Video

Future Scope

We had a good run working on this project but there is still scope for improvement. We can start with incorporating the feedback given by the users like adding profile customizations and adding an option to hide annotations, then we can work on improvising the security and the UI of our product.

The Team

We would like to thank Prof. Ponnurangam Kumaruguru for his continuous guidance in this project. Working on this project was an amazing experience for all of us. We hope that NewsBuddy would be able to make a real-world impact! :D