Defining the Hate Speech Crime

By Nanjira Sambuli
iHub Research
  Published 20 Nov 2015
Share this Article

By Kagonya Awori, Umati Project Researcher.

On October 12th 2015, political analyst Mutahi Ngunyi was charged with four counts of ethnic contempt and hate speech, due to statements he had posted on his Twitter account between 19th and 24th August 2015, that were purportedly discriminatory.

The case was presented to the National Cohesion and Integration Commission (NCIC) by the Law Society of Kenya CEO, Apollo Mboya, who was himself a target of Ngunyi’s reportedly hateful tweets. Ngunyi’s is not the only high profile case on hate speech in recent years. In 2010, Wilfred Machage and others were arraigned in court over utterances made during campaigns against the constitutional referendum. In 2012, Ferdinand Waititu made utterances that allegedly led to the death of two Maasai men in Kayole. His case is still ongoing. In 2011, Chirau Ali Mwakwere was cleared of hate speech charges brought against him due to utterances he allegedly made in July 2010, claiming that Arabs had oppressed indigenous coastal people. More recently, Gatundu South MP Moses Kuria was charged with incitement and hate speech over posts he made on his social media pages allegedly against the Luo community. The case is still in court.

None of these cases involving politicians has resulted in the up to 5 year jail term or the Kshs. 1 million fine that the 2008 National Cohesion and Integration (NCI) Act (Section 62) stipulates as punishment for offences of ethnic and racial contempt. Contrarily, most of these cases have been quashed after the accused has offered a mere apology to the NCIC.

On the flip side, university student Allan Wadi was in January this year charged with hate speech over statements he made on his Facebook page alleging that a particular tribe should be deported from Kenya, and critiquing President Kenyatta. He was sentenced to two years in jail.

These and other hate speech cases raise concern over how seemingly flexible the hate speech charge is, depending on who has been charged, who brings the charges, the relationship between the accuser and the accused, and whose name(s) has been mentioned in the hate speech statements. While these four factors seem to be the criteria for determining what is or what is not hate speech vis-a-vis the cases above, officially, they are not.

According to section 13 of the 2008 NCI Act, a person who uses, displays, publishes, distributes, shows or plays threatening, abusive or insulting words or behaviour (including written material, programs, visual images, recordings or plays),  “commits an offence if such person intends thereby to stir up ethnic hatred, or having regard to all the circumstances, ethnic hatred is likely to be stirred up.”

Ethnic hatred is further defined as “hatred against a group of persons defined by reference to colour, race, nationality (including citizenship) or ethnic or national origins.”

Interpreting the NCIC definition above, hate speech is committed when a person uses threatening, abusive or insulting words, texts, images or acts to stir up hatred against a group of persons based on their race, nationality, citizenship, ethnicity or national origin.

The criteria here includes the intent, the content of the speech act, and the basis of the harm i.e. race, ethnicity etc.

The question that is raised then is, what definition of hate speech does the mwananchi live by? The definition implied by cases like Wadi’s and Ngunyi’s where personal insults to politicians are apparently hate speech, or the definition provided by the NCI Act? Are we freely engaging within our freedom of speech, or are we stifling potentially fruitful political discourse in the fear of committing the hate speech crime?

In 2012, our hate speech monitoring project, Umati, sought to understand what the average Kenyan considers hate speech. The Umati project was launched in October 2012, to collect and analyse online hate speech statements around the 2013 general elections. The project was motivated by the 2007/8 post election violence whichhighlighted the potential for speech to catalyse mass violence on the ground and the role that mobile and digital technologies may have played in propagating such speech. The Umati project is still ongoing, with an added focus on building an intelligent tool that augments thecollection and monitoring of online hate speech.  Umati’s intent is not to find and report hate speech offenders, but instead explore non-punitive, citizen-centred approaches to mitigate hate speech online.

As a starting point, and in order to standardise what could be considered “threatening, abusive or insulting words or behaviour” to “a group of persons”, Umati extended the NCIC definition of hate speech to include dangerous speech - speech that has a high potential to catalyse violence. The dangerous speech framework is based on extensive research on cases where vitriolic speech resulted in extensive harm, including during the genocide in Rwanda and the 2007/8 Post Election Violence in Kenya. We adopted aspects of the dangerous speech framework that would standardise how Umati identifies, and thus monitors, hate speech. For example, we included discrimination that occurs along more socio-cultural lines eg gender, sexual orientation, religion, political affiliation etc, and not only along racial, national or ethnic lines as stated in the NCI Act. We also attempted to be more categorical and specific about what may constitute “threatening, abusive or insulting” words or acts. Thus, we adopted what the dangerous speech framework proposes are key characteristics common across speech acts, that in the past resulted in grave violence against groups of people.

Therefore, the definition of hate speech that the Umati project adopted is speech that,

  1. 1. is targeted at a group of people and not a single person.


  1. 2. may do one of the following:
  • Compare a group of people with animals, insects or vermin
  • Suggest that the audience faces a serious threat or violence from another group and thus needs to arm themselves or attack first
  • Suggest that some people from another group are spoiling the purity or integrity of the speakers’ group


  1. 3. contains a call to violent action, that is a call to discriminate, loot, riot, beat, forcefully evict, or kill a group of people, or a person because of their belonging to a particular group.

It was necessary to expressly stipulate that hate speech does not constitute “threatening, abusive or insulting words or behaviour” against a single person, but it is harmful speech against a group of people (as also stated in the NCI Act), or against a person because of their belonging to that particular group of people.

However, while the definition we adopted is clear, specific, applicable and easy to understand, it is not the definition the public necessarily follows, nor that NCI Act stipulated.

We thus conducted a small pilot study to investigate what the Kenyan public understood as hate speech, and how their perception compares to the Umati definition and the NCIC definition. We conducted the study with 12 participants, where we presented them with 190 comments collected from Kenyan Facebook pages. Participants were asked to go through the Facebook comments and identify which of the comments constituted hate speech, and what severity rating they would assign to each hate speech statement. The rating was between 1 and 3: a rating of 1 indicated that the comment was ‘barely inflammatory’ and a rating of 3 indicated that the comment was ‘extremely inflammatory’. For comparison, we asked 6 Umati project staff to take part in the study. The staff were all trained in identifying hate speech according to the aforementioned framework.

From the 190 Facebook comments, the Umati staff identified 44 statements as hate speech whereas the public identified 156 statements as hate speech. Further, only 5% of the statements (10 out of the total 190) were picked by both the Umati staffand the public. In other words, out of the 190 statements, for only 10 did both the Umati staff and the public agree that they were hate speech statements.

The results from this study revealed that the public have a broad definition of hate speech. The participants considered personal insults, critiques of government officials, propaganda and rumours as hate speech. While propaganda for war and discrimination are outside one’s freedom of expression, they are not hate speech. Hate speech is a criminal offence, libel and defamation are not. The participants' broad inclusion of all negative commentary on politicians as hate speech, can be seen as an effect of cases such Wadi’s and Ngunyi’s, where it appears that the accused was charged due to painting particular politicians in a negative light. It needs to be clear to the public which statements made by offenders are hate speech, in order that political speech - which is legal and protected - is not confused with hate speech. As noted by Dr. Onyoyo from the Nairobi School of Law:

“Is writing and publication of a book tainting the personality of a person a hate speech? The answer is no. Such offences under tort which can be filed under civil proceedings as libel or defamation and not as hate speech. The example is the publication of the book ―Peeling Back the Mask by Miguna Miguna that damaged the political image of Hon. Raila A. Odinga, the then Prime Minister of the Republic of Kenya could be categorized under civil offence (libel) but not criminal offence (hate speech). The book circulated widely even online and in national media but no legal action was taken against the author by the offended.”

Thus, while the line between hate speech, incitement to violence, libel and defamation is arguably a grey one especially in our ethnically charged socio-political sphere, it is still a line that can be made clearer. Umati attempted this by appropriating the dangerous speech framework, whichclearly defines a type of speech that has a high potential to cause harm to a group of people. Clear definitions counter ambiguity and facilitate harmony. The clear definition Umati adopted allowed the project to identify which statements are hate speech, and which, though negative or insulting, are not hate speech.

While we are not necessarily proposing that the definition we used be adopted nationally, our small study shows that it is possible to make the grey line clearer. The framework we adopted can be learnt and applied easily and thus allow the public, as exemplified by the Umati staff in our study, to clearly identify, and know how to respond tohate speech. With the 2017 elections coming up, it is critical to equip the public with knowledge on how to detect, respond to and avoid hate speech so that they can be immune to politicians who unabashedly promote conflict through speech.

Relatedly, the public’s broad understanding of hate speech may hamper their freedom to engage in political speech. With cases such as those of Allan Wadi apparently setting precedent, the mwananchi may be deterred from justly critiquing actions or inactions of the political class, in fear that they too may be prosecuted.

In conclusion, the definition of hate speech is a contentious issue in most, if not all countries around the world. In fact, Kenya’s attempt to legally define it and prosecute it is highly commendable. However, there is still a real need for consistency across defining, prosecuting and curbing the ‘hate speech crime’, especially among the political class. The current unspecific definition of hate speech in Kenya has thus far only officially made things worse; the unabashed political class who use it to promote mass violence escape through the cracks, while the less affluent mwananchi is punished. Instead, avenues should be sought that protect and sensitise the public on being immune to hate speech, and fairly persecute those who commit hate speech. However, this cannot happen if the definition of hate speech is vague, or malleable depending on the accused’s bank account and political influence. Thus, a first step, as we did in Umati, is to implement a clear, workable framework for defining, identifying and responding to hate speech, in order that peace and harmony is upheld in both online and offline political fora. This can in turn help in understanding and upholding freedom of expression as enshrined in the constitution, a right that every Kenyan should enjoy, and a key instrument in fostering cohesion in our country.


(While Umati is defined as a dangerous speech project, we use the terms dangerous speech and hate speech interchangeably. We further propose that the definition we adopted for dangerous speech can be considered as a definition for hate speech given that it is more specific and actionable than the current vague definition.)

comments powered by Disqus