Read time: 3 minutes

Kenya’s online hate speech ‘spiked’ before March elections

Kenya , 15 Apr 2013

Kenya’s online hate speech ‘spiked’ before March elections

Levels of online hate speech in Kenya increased just before the country’s March 2013 elections, says a report from Nairobi technology centre iHub and crowdsourcing company Ushahidi.

Researchers of the ‘Umati: Monitoring Online Dangerous Speech’ report, which was released last week, say they started collecting data in September 2012. Since that time, they have monitored and recorded incidents of hate speech and categorised it into classes regarding order of severity.

Categories include offensive speech, comments meant to insult a particular group, moderately dangerous speech, comments which are moderately offensive, ‘extremely dangerous speech’, comments which are ‘extremely inflammatory’ and those comments that have the highest potential to incite violence.

Unlike the Kenyan government, which has reportedly been using software to monitor online hate speech, the Umati project has used human monitors to screen public posts on Facebook, Twitter and blogs.

Angela Crandall, research manager at iHub, said that the researchers adopted this approach because hate speech could take on various vernaculars, a phenomenon that software may fail to interpret

And the Umati report goes on to highlight that the number of offensive online speech instances recorded in Kenya amounted to 405 in March, a dramatic rise from 122 in February. Meanwhile, extremely dangerous speech instances rose from 197 in February to 321 in March.

Awareness of hate speech in Kenya has been on a high in the country following violent elections in 2007 and 2008. At the time, a disputed presidential election result between front-runners Mwai Kibaki and Raila Odinga sparked ethnic violence and tensions within the country.

“Between November 2012 and January 2013, Umati saw a declining number of extremely dangerous speech statements (category 3 statements) on the Kenyan online space,” says the report.

“However, in the weeks leading up and following the elections, we see an increase in dangerous speech,” adds the report.

Findings of the report further indicate that Facebook is where most online hate speech conversations occur in Kenya, both in public and private groups.

Umati reports online speech, which is deemed to be potentially harmful to society, to the police or the National Cohesion and Integration Commission (NCIC), says Crandall.

Researchers of Umati are planning to release a final report in June this year.

Daily newsletter