Acerca de

​

 

 

 

 

 

 

 

 

For the second year in a row, Bunka's technology has been employed to craft a part of the annual report for the Human Rights Council's committee on hate speech, in collaboration with the Medialab of Science Po.

 

The main goal was to assess and quantify various forms of online hate. To achieve this, we collected over 2,000,000 YouTube comments and utilized Bunka's modules to gain a comprehensive perspective on different categories of hate speech, including Islamophobia, anti-Semitism, and racism.

​

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

These visual representations, or maps, are generated by organizing comments into a two-dimensional space using a range of LLM algorithms. The denser a particular area, the higher the concentration of comments within it. Labels serve to encapsulate specific map regions, with each label summarizing thousands of comments and assigning a descriptive tag. Notably, two closely positioned labels on the map signify a stronger resemblance than those situated farther apart, in line with Tobler's first law of geography. These map "cards" portray thematic maps, wherein the proximity of two themes indicates their shared characteristics.

​

The full report (In French) can be found here.

 

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​