AI developed to help police support victims of hate crime

Their results show that as the number of “hate tweets” – those deemed to be antagonistic in terms of race, ethnicity or religion - made from one location increased, so did the number of racially and religiously aggravated crimes – which included violence, harassment and criminal damage.

Researchers add that an algorithm based on their methods could help police predict and prevent spikes in crimes against minorities by allocating more resources to specific areas.  

Director of HateLab Professor Matthew Williams said: “This is the first UK study to demonstrate a consistent link between Twitter hate speech targeting race and religion and racially and religiously aggravated offences that happen offline.

“Previous research has already established that major events can act as triggers for hate acts. But our analysis confirms this association is present even in the absence of such events.

“The research shows that online hate victimization is part of a wider process of harm that can begin on social media and then migrate to the physical world.”

Computer scientists developed artificial intelligence to find 294,361 “hateful” Twitter posts during an eight-month period between August 2013 and August 2014. A total of 6,572 racially and religiously aggravated crimes were also filtered out of police data. These figures, along with census data, were then placed into one of 4,720 geographical areas within London to allow researchers to pinpoint trends.

Professor Williams, who is based in the School of Social Sciences, added: “Until recently, the seriousness of online hate speech has not been fully recognised. These statistics prove that activities which unfold in the virtual world should not be ignored.

“The data used in this study were collected at a time before the social media giants introduced strict hate speech policies. But rather than disappear, we would expect hate speech to be displaced to more underground platforms. In time, our data science solutions will allow us to follow the hate wherever it goes.”

HateLab is a global hub for data and insight into hate speech and crime. Using data science methods, including ethical forms of AI, the initiative was set up to measure and counter the problem of hate both online and offline.

The Online Hate Speech Dashboard has been developed by academics with policy partners to pre-empt outbreaks of hate crime on the streets. It has been established with funding from the Economic and Social Research Council (ESRC) as well as the US Department of Justice. It has received a total of £1,726,841 in funding over five ongoing projects.