The growing prominence of fake news since the US presidential election last year has made it the talk of the town – with major companies like Facebook actively trying to curb the spread of erroneous news circulating online.
So when 140 students were tasked to come up with tech solutions to combat fake news at a hackathon over the weekend, one group in particular stood out in the eyes of the judges with its “three-pronged” approach.
The hackathon, organised by the National University of Singapore (NUS), Singapore University of Technology and Design (SUTD), Google and the Media Literacy Council, was held on Oct 14 at the Google Asia-Pacific office in Pasir Panjang.
The winning team, a group of five first-year students from the Singapore University of Technology and Design, dubbed themselves “The Losing Team” because they didn’t expect to win, said team member Mr Brandon Ong, 21.
He said: “We just joined the competition to learn and experience as much as possible”.
Ironically, the odds were ever in their favour.
In a telephone interview with Business Insider, Mr Ong explained his team’s “three-pronged” idea:
A web-based crawler is used to search and analyse the credibility of online articles, with the information collected stored on an online database, explained Mr Ong.
For the purpose of the hackathon, the team used a cloud server called DigitalOcean, but they hope to seek out an alternative server in the long run.
The Google Chrome extension
After an online database is created, users will be able to download a Google Chrome extension. When the extension is activated, each article viewed will be given a percentage rating (from 0% to 100%) evaluating its credibility.
This percentage rating will be based on a sentiment analysis conducted using natural language processing.
In other words, the objectivity of language used in each article will be taken into consideration when evaluating its credibility.
“In the future, we can also include the credibility of the news source itself and information from our crowd sourcing platform. But we haven’t started on that yet”, said Ong.
A community crowd sourcing platform
On a platform similiar to Reddit or Quora, users will be able to rank the credibility of articles that they read (on a scale of one to 10).
A higher score will reflect a higher level of credibility.
“By getting people to come together, we can encourage discussion and use that information to check for credibility”, said Ong.
He adds that while the team is happy about winning the hackathon, it “would be cool if Google picks up (their idea)” so that they can start working on a more concrete plan.