Home AI Ethics Computational models to detect stereotyping of ethnic minorities in News

Computational models to detect stereotyping of ethnic minorities in News

Various studies have revealed that conventional news sources often create stereotypical or discriminatory impressions of ethnic minorities. In Norway, surveys conducted on Muslims and non-Muslims reveal that the Norwegian population sees news media as a significant source for this negative impression. However, to recent date, data-scientific research on how mainstream Norwegian news media writes about these minority groups was still absent.

Laréb Fatima Ahmad investigated, using Natural Language Processing methods, if stereotyping of Muslims and ethnic minorities could be objectively detected in newspapers. For her research Fatima used as input data, newspapers from 2017 to 2019. The study — which could only be done with computational liguistic methods, given the large body of text — objectively shows a disproportionate reportage practice in Norwegian newspapers in which Muslims and ethnic minorities are described in ways that align with bias towards negative, discriminatory stereotypes.

Lareb Fatima Ahmad (2024) The Islamic Elephant in the News Room
Using Natural Language Processing to Uncover Stereotypes of Muslims and Ethnic
Minorities in Norwegian News in 2017-2019. Master thesis Computer Science, Vrije Universiteit Amsterdam. The thesis is available at request.