Bibliographic citations
Mostajo, A., (2024). Sesgos de la inteligencia artificial en la representación visual de los pueblos indígenas amazónicos [Tesis, Universidad Peruana de Ciencias Aplicadas (UPC)]. http://hdl.handle.net/10757/673913
Mostajo, A., Sesgos de la inteligencia artificial en la representación visual de los pueblos indígenas amazónicos [Tesis]. PE: Universidad Peruana de Ciencias Aplicadas (UPC); 2024. http://hdl.handle.net/10757/673913
@misc{renati/417894,
title = "Sesgos de la inteligencia artificial en la representación visual de los pueblos indígenas amazónicos",
author = "Mostajo Pantoja, Augusto Franco",
publisher = "Universidad Peruana de Ciencias Aplicadas (UPC)",
year = "2024"
}
The research aims to analyze the images generated by two Artificial Intelligence platforms, DALL-E and Runway, on the representation of Amazonian indigenous peoples in order to identify biases in them. A sample of 400 images were analyzed in the fields referred to the number of people, gender, age range, geographic space where they are located and the characteristics in the composition of the images. The prompt used to generate the images was: Amazonian indigenous people. On the one hand, DALL-E represented Amazonian indigenous people as individual subjects in 69.5% and omitted their presence in 27% of the images. On the other hand, Runway showed a clear tendency to represent only male Amazonian indigenous people in 69% of the images. In addition, more than half of the images on both platforms presented images in which the geographic space could not be determined and none of the images depicted an urban environment. These results show biases in the generation of images, which derive from the way in which the Artificial Intelligence systems have been built or are being trained.
This item is licensed under a Creative Commons License