Social media tool could anticipate real world hate crimes


Academics from Cardiff University’s HateLab initiative have said that an algorithm based on their methods could be used to predict crimes against minorities in specific areas.

Social media tool could anticipate real world hate crimes
Social media tool could anticipate real world hate crimes
The statement follows an analysis by the university over an eight month period of Twitter activity, correlated with recorded crime data from the Metropolitan Police. The results show a significant link between the number of ‘hate tweets’ made from a particular location, and the level of “racially and religiously aggravated” crimes taking place in the real world.  
Speaking of the findings, 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.
“The data used in this study was collected at a time before the social media giants introduced strict hate speech policies. Rather than disappear however, 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.

“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.”
As part of the project, computer scientists developed artificial intelligence through which were found nearly 300,000 ‘hateful’ Twitter posts between August 2013 and 2014. The figures - along with census data - were subsequently placed into one of 4,720 geographical areas within London.

To hear the latest on digital communications best practice, register your interest for and visit BAPCO 2020


Philip Mason
Editor, Critical Communications Portfolio
Tel: +44 (0)20 3874 9216

Author: Philip Mason