Technology experts in Qatar and India have developed a new online application that rates the credibility of Twitter feeds, in a bid to stop people from spreading rumors online – particularly during major incidents.
TweetCred has been thought up by developers at Qatar Computing Research Institute (QCRI) – a private, not-for-profit national research organization founded by Qatar Foundation – and Indraprastha Institute of Information Technology (IIIT) Delhi, which has been collaborating on the project for the past six months.
The aim to to give people a way of sorting through their Twitter feed quickly, to help them identify credible information sources in times of national incidents and emergencies such as riots, public uprisings and earthquakes.
The new app, which can be downloaded for free as a Google Chrome extension, comes as Twitter becomes one of the main sources of breaking news and information for people in a crisis.
In Doha, for example, some 90 percent of residents rely on the internet for news, according to a 2013 Northwestern University in Qatar study on media usage in the MENA region. And some 43 percent of people surveyed in the country use Twitter.
TweetCred works by giving each tweet a rating, from one to seven, according to how credible the user posting the message is believed to be. The greater the number of blue dots, the more credible the tweet.
The system uses algorithms that take into account a number of factors, including whether any pictures or video are included in the tweet, as well as the reputation and popularity of the user who posted the message.
In his technology blog iRevolution, QCRI’s Director of Social Innovation, Dr. Patrick Meier, explained that the app puts to a practical use academic information his colleagues at QCRI and IIITD gained during previous research.
That information includes an automatic analysis of 35 million tweets about a dozen major events such as the London riots of 2011, the Chilean earthquake of 2010 and the uprising in Libya.
Speaking to Doha News about the app’s purpose, Meier said:
“TweetCred is more important for the behavior change that may result than the actual technology itself. By simply adding a credibility score, we hope that Twitter users will think more critically about what they read and retweet on Twitter.
We want a more informed and critical digital public sphere.”
He added that such a system is particularly useful in emergency situations, where it is crucial for people to find out information quickly and accurately.
He said the hope is to also help the Twitter community at large to become more self-regulating, allowing them to easily see who is a reliable source and who is not.
“Studies have shown that we’re less likely to spread rumors on Twitter if false tweets are publicly identified by Twitter users as being non-credible,” he said.
Additionally, public exposure increases the number of Twitter users who seek to stop the spread of rumor-related tweets by 150 percent, Meier added.
The app is still in an experimental stage, and the developers describe it as a “hybrid,” combining big data with personal input.
To help refine it, they are asking users to click on a thumbs-up or thumbs-down icon to agree with the rating given to a particular tweet. If they disagree, users are asked to change the credibility rating.
As the system is powered by “machine learning,” it becomes smarter as more people interact with the system and teach it how to accurately rate tweets.