Researchers developing a social media lie detector

20 Feb 2014

Image via artiomp/Shutterstock

An international group of researchers led by the University of Sheffield are trying to find a way to verify information spreading on social networks in real-time, in order to provide a reliable and valuable resource for journalists, governments, emergency services and health agencies.

In Greek mythology, Pheme would tell tales first with a whisper but then gradually getting louder until word had spread to everyone. This is precisely how rumour and misinformation can spread on social media and is the same name given to an EU-funded project researching how to sort the truth from the lies.

Researchers from seven different countries are taking a multidisciplinary approach to the problem, combining big data analytics with natural language processing and text mining, web science, social network analysis, and information visualisation.

How Pheme works

Resources testing the veracity of user-generated content already exist. For example, Mark Little’s Storyful (recently snapped up by News Corp for €18m) uses teams of humans in Asia, Europe and the US working around the clock to source and verify news breaking on social networks.

This manual process is time-consuming and requires significant resources, whereas Pheme uses a computer system that, essentially, acts like a lie detector and provides instant results.

Using the WebLyzard web intelligence platform, Pheme identifies and tracks four categories for online rumours: speculation, controversy, misinformation and disinformation (the difference between the latter two being that one is spread unwittingly and the other is done with malicious intent).

Sources will also be categorised to determine their authority, giving more weight to established news outlets and experts and ignoring bots sending out spam.

Pheme will then search for other sources to corroborate or deny information and plot how conversations on social networks evolve, using all of this information to determine if the claim is true or false. The results are then displayed on-screen, telling users if lies are taking hold.

Stopping the spread of bad information

A service like this has many uses. During the 2011 riots in London, claims that landmarks had been set alight and animals had been released from the zoo diverted emergency services. Much incorrect reporting surrounded the Boston Marathon bombing last year, fuelled by the rapid spread of incorrect information online. Not to mention the dangerous and damaging rumours that can be spread during global conflicts.

The Pheme research project is funded by the European Commission within the 7th Framework Programme (FP7) and has a three-year life-span. It’s expected that a final version of the product could be completed within 18 months, while working prototypes could be available sooner.

Truth or lie image by artiomp via Shutterstock

Elaine Burke is the host of For Tech’s Sake, a co-production from Silicon Republic and The HeadStuff Podcast Network. She was previously the editor of Silicon Republic.

editorial@siliconrepublic.com