Htqe.2025.23

https://doi.org/10.22364/htqe.2025.23 | 327-339 | PDF

Bias-Related Ethical Considerations of Generative Text-to-Image (TTI) Technology: Educational Perspective

Eva Strazdiņa¹
¹University of Latvia, Latvia

Abstract
Generative text-to-image (TTI) models hold significant potential across various fields. Their widespread recognition stems from their ability to produce high-quality imagery based on textual descriptions. Despite the vast range of possibilities, ethical concerns and risks associated with computer-generated imagery remain a significant point of discussion. While the European Union Artificial Intelligence Act provides a legal framework for the development and use of general-purpose AI systems, including TTI technology, it is essential to assess the potential risks associated with their application in education and vocational training. Applying TTI technology in producing imagery may result in images that reinforce stereotypes, perpetuate social biases, and limit the representation of diverse cultural aesthetics. To analyze and identify the characteristics of such biases and bias-mitigating strategies, this study poses the following research question: 1) What bias-related ethical considerations relevant to educational contexts emerge in using generative text-to-image (TTI) technology? A scoping review was employed as the research method. The Web of Science and Scopus databases were chosen, and after conducting a content analysis of 225 abstracts of open-access articles in English, published between 2019 and 2024, containing the keywords “text-to-image” and “bias”, 24 articles were selected as relevant to the focus of this scoping review. ​​From the analysis of the studies, three main categories emerged: (1) Characterization of bias, (2) Bias-mitigating strategies, and (3) Bias-related ethical considerations within an educational process. These categories were further broken down into more specific bias-related ethical considerations and their descriptions. This research contributes to the field by highlighting bias-related ethical considerations that educational professionals should know before incorporating TTI technology into education and vocational training.

Keywords: text-to-image (TTI), generative AI, bias, ethical considerations, education, scoping review

In: Human, Technologies and Quality of Education, 2025. Proceedings of Scientific Papers = Cilvēks, tehnoloģijas un izglītības kvalitāte, 2025. Rakstu krājums.  Ed. L. Daniela. Riga: University of Latvia, 2025. 698 p.
ISBN 978-9934-36-461-7 
https://doi.org/10.22364/htqe.2025