New project to check hate speech on Facebook, Twitter
Posted Sunday, February 3 2013 at 00:30
- A group is on call 24 hours to filter updates on the social media
Be careful what you tweet, update on your Facebook account or write on your blog.
An online monitoring project has been trawling websites, blogs and the social media for hate speech in the past four months ahead of the March 4, 2013 elections.
The Umati project, an initiative of technology incubator, iHub, and software developer, Ushahidi, aims to monitor and report dangerous postings online by Kenyans to the authorities.
Ms Angela Crandall, the research manager at iHub, said five monitors are at the firm’s research centre.
“We have people who in addition to mastery of English understand Dholuo, Kalenjin, Kikuyu, Luhya, Swahili and Sheng,” said Ms Crandall, who has been working at iHub since 2011.
“We once toyed with the idea of using computer software, but much content was in vernacular and it was difficult to develop software to interpret this.”
The monitors follow opinion and political leaders from different communities and collect data from Twitter, Facebook, blog posts and comments on online newspapers.
The data is then categorised as offensive speech (lowest level), moderately dangerous speech or extremely dangerous speech (highest level).
Ms Crandall told the Sunday Nation that Umati shares the reports it develops with different partners including the police, National Cohesion and Integration Commission and the civil society. They are also in contact with the ministry of Communication as well as electoral commission.
These sites have gained massive popularity since the last General Election, a period when radio stations and mobile phones were the main mediums used to perpetrate hatred.
In contrast to 2007 when there was no systematic monitoring of the Kenyan online environment, Umati is trying to ensure that it captures the trending topics, phrases, and sentiments especially in this campaign period.
“We are aware of the influence that new media had on the 2008 post-election violence in Kenya and this is why we are undertaking this project,” says Crandall.
But why “dangerous speech” and not “hate speech?”
The National Cohesion and Integration Commission Act of 2008 says hate speech is “speech that advocates or encourages violent acts against a specific group, and creates a climate of hate or prejudice which may, in turn, foster the commission of hate crimes”.
The Umati monitors, however, particularly go after “dangerous speech” or “speech that has a potential to cause violence”.
“The definition that NCIC has for hate speech is vague and leaves room for different interpretations,” said Ms Crandall.
“Dangerous speech” is borrowed from Prof Susan Benesch, an academic who studies the role of inflammatory speech in catalysing collective violence around the world.
Ms Crandall said that inflammatory speech is still rife in the Kenyan online space.