10 November 2020

What forms of hate speech will artificial intelligence find online? What are the channels where hate speech flourishes? How does hate speech differ from channel to channel and platform to platform? These are some of the questions to be addressed in the project that Finland’s Ministry of Justice has started with the help of Utopia Analytics. The findings will be used for the Finnish government’s work combating hate speech.

This project is part of The Facts Against Hate program which is coordinated through the Ministry of Justice, and the partners are the Ministry of the Interior, the Police University College, the Anti-Racist Forum, the Centre for Peace Studies (Croatia), and INAR (Ireland). The program will include a project in which a tool based on artificial intelligence is built by Utopia Analytics.

The goal of the project is to examine the extent of hate speech. It will study, for instance, which channels most often include hate speech, the themes of hate speech, and the ways that hate speech is affected by real-world events.

Working in cooperation with Ministry of Justice experts, a university researcher will define what is to be considered hate speech messaging. Utopia will train Utopia AI to find content similar to this definition online.

The monitoring will focus on content from Twitter, Facebook, Instagram, discussion forums, the news comments on major news sites, blog posts and blog post comments. The data is purchased from a third party for the project. This approach will allow overall comparisons among all the most relevant and popular platforms. The aim is to create a big picture of hate speech, not to evaluate the content of individual messages.

The findings from this monitoring process will be used in the government’s work against hate speech. The results will be published during the first half of 2021.

“For years now,” says Utopia’s CEO Mari-Sanna Paukkeri, “many operators in Finland and worldwide have used Utopia’s technology in detecting and removing inappropriate content from their online services. In this project we dive deep into hate speech in particular. It’s very interesting to compare the appearance of hate speech on different platforms.”