Scientists are developing a Lie-Detector for twitter

Researchers at the University of Sheffield have been given an EU grant to develop an automated system to grade the reliability of online assertions and rumours, so we all know what to believe. Five European universities, led by Sheffield in northern England, are cooperating on a system that could automatically identify whether a rumour originates from a reliable source and can be verified.

The ambition for a computerised system that will be able to sort truth from lies on Twitter and elsewhere is certainly in tune with our modern algorithm fetish, the curious assumption that software should replace human judgment wherever possible. But algorithms are designed by humans, and their workings depend on the human assumptions that were baked into them. So what are the assumptions governing the social media lie detector?

The three-year, European Union-funded project, called PHEME, is an attempt to filter out the nuggets of factual information from the avalanche of ill-informed comment on Twitter and Facebook. Under the project, set to take at least 18 months to complete, falsified tweets will be grouped into one of four categories: speculation, controversy, misinformation and disinformation.

The research is being led by Dr Kalina Bontcheva of Sheffield University's Faculty of Engineering.

"The problem is that it all happens so fast and we can't quickly sort truth from lies," she said. 

"This makes it difficult to respond to rumours, for example, for the emergency services to quash a lie in order to keep a situation calm. Our system aims to help with that, by tracking and verifying information in real time."

The researchers say they hope the system will allow governments, emergency services, media and the private sector to respond more effectively to claims emerging and spreading on social media before they get out of hand.