The anti-social network: tracking the trolls

Reblogged from Lancaster University’s STEPS magazine, Summer 2015, p13-14.

A team from Lancaster University led by Dr Claire Hardaker has produced a cutting-edge tool that could help to pinpoint cyberbullies, trolls, and even radicalising groups on busy social networks such as Twitter. Dr Hardaker, a lecturer in forensic corpus linguistics, began an ESRC-funded project researching abusive behavior on Twitter in December 2013.

The project quickly demonstrated that whilst tackling antisocial online behavior is of key importance, sites like Twitter produce data at such high volumes that simply trying to identify relevant messages in amongst all the irrelevant ones is a huge challenge by itself. However, less than a year into the project, Hardaker and her team were invited to Twitter’s London headquarters to present findings from the project to Police, the Crown Prosecution Service, and Twitter itself, and the results of this research subsequently encouraged Twitter to update its policy on abusive online behavior. The interest from the police and Crown Prosecution Service also encouraged Hardaker to turn her research into a tool that could filter the ‘noise’ out of billions of online messages, thereby enhancing the useful ‘signals’ that can lead to the identification of accounts, texts, and behaviours of interest. This software – known as FireAnt – was built for Hardaker by fellow corpus linguist, Professor Laurence Anthony, and whilst initially designed to handle data from Twitter, it can also analyse data from almost any online source, including sources such as Facebook and Reddit.

Hardaker said:

The Government is trying to understand how social networks are involved in all sorts of issues like child-grooming, human trafficking, and radicalisation. FireAnt offers the ability to analyse huge datasets to identify abusers, their networks, and their targets.

A key aspect of this analysis in Hardaker’s work is the process of escalation from online behavior that is simply unpleasant or annoying up to extreme, illegal behavior, and even that which could turn into physical, offline violence. One potential use for FireAnt is to pin-point high-risk individuals and networks who may go on to be a threat, whether to themselves or others.

Hardaker’s knowledge of the English language has already proved crucial at a national level. Her skills were used in 2014 to analyse videos to try to identify the Islamic State (Isis) militant with a British accent who appeared in a video of the apparent beheading of US journalist James Foley. Hardaker studied the clip and said the man’s vowels marked him out as probably from the south east of England.

Hardaker specialises in aggression, deception, and manipulation online. She is also working on other projects that involve analyzing live online social networks for escalation of abusive behavior, the importance of the internet in transnational crime such as human trafficking, and the internet as an arena for grooming and coercing victims.

Originally from Bradford, she took her Masters and PhD at Lancaster University. Initially interested in aggressive online behavior, Hardaker started studying trolling when she realized that there was virtually no research on it, and what started as a simple interest has now proved to have implications for national and international security. She has featured extensively in the media in interviews where she is regularly consulted for her breadth of knowledge of her research.