Junk news aggregator aims to restore trust in media and democracy
08 May 2019


Developed by the Oxford Internet Institute with EU funding, the junk news aggregator (JNA) interactively displays articles from unreliable sources as they spread on Facebook. Researchers hope the tool will help tackle the growing phenomenon of misinformation on social media.

Cover image of Junk news aggregator aims to restore trust in media and democracy

Named Word of the Year by Collins in 2017, ‘fake news’ has slowly moved away from sensationalist and populist discourse to anchor itself in our everyday lives. Fake news, or some would rather say junk news, is now everywhere, starting with our own Facebook walls. They’ve grown more credible, too: Unless you’ve been educated to second-guess every article spreading across your network, and have the time to cross-source information, the odds of getting caught up in the fake news frenzy have never been so high.

Introducing the JNA

Luckily for all of us, a group of scientists at the University of Oxford have taken an important step in the systematic evaluation of news credibility on social media. Their JNA was launched in the lead-up to the 2018 US midterm elections. It can track down junk news sources on Facebook and aggregate them on a dedicated online portal, so that none of us can get fooled again.

“Earlier attempts to address misinformation on social media often focused on Twitter only. They did not offer real-time insights into currently popular misinformation and junk news content on social media, nor did they allow the public to examine, filter or search through this junk news content as it spread on social media,” says Dr Dimitra Liotsiou, one of the researchers leading the BOTFIND (Finding Bots, Detect Harassing Automation, and Restoring Trust in Social Media Civic Engagement) project. Her aggregator, however, does it all.

Three distinct tools

The platform consists of three distinct tools: an exhaustive list of posts by junk news sources and their content posted publicly on Facebook, with filters by release date, engagement level and keyword; an interactive visual aggregator of the day’s most popular posts by junk news sources; and a ‘Top 10’ snapshot of the day’s most popular posts by these same sources.

As Dr Liotsiou explains: “The aim of this public tool is to make the issue of junk news on social media more transparent, while enabling journalists, civil society groups and all interested members of the public to examine in real time what kinds of junk news are spreading on social media. This helps raise awareness and improve media literacy, and should ultimately contribute towards preventing users from being influenced by online misinformation and junk news.”

So how exactly does the aggregator work? In the case of the 2018 elections, the team started by identifying tweets referencing these elections on Twitter and extracting included links. From thereon, they categorised the source of each link as junk news as soon as it failed on three out of five criteria: whether the source follows professional journalistic standards; whether fact-checking is done; whether commentary is disguised as news; whether reporting is highly biased, ideologically skewed or hyper-partisan; and whether the source counterfeits the branding of other established news outlets.

“For each of the top 50 most cited (on Twitter) junk news source websites, all the posts each source uploads onto its public Facebook page are retrieved and displayed on the main JNA tool every hour. The tool displays the content, image or video, link, and all engagement numbers for each post,” Dr Liotsiou explains.

By helping shed light on junk news on social media, the team hopes that the JNA will contribute to restoring public trust in technology and the process of modern deliberative democracy. The tools are open and publicly available, and can help journalists, public policy makers, civil society leaders, politicians and members of the public to access, examine and evaluate news quality online in a timely manner.

Project information

BOTFIND: Finding Bots, Detect Harassing Automation, and Restoring Trust in Social Media Civic Engagement
Phil Howard
Host institution:
The Chancellor, Masters and Scholars of the University of Oxford
United Kingdom
Call details
ERC Funding
149 921 €