Abstract—Limited or superficial knowledge about others can foster stereotypes and prejudice. Consequently, this study explores specific methods to counteract these stereotypes. We posit that the key to challenging stereotypes lies in acquiring relevant knowledge about the subject. Given that gathering such information is often time-consuming, this study introduces an Artificial Intelligence (AI) -assisted, time-efficient approach. Additionally, we outline the advantages and disadvantages of these methods and examine their suitability across different scenarios. This study has three primary objectives. First, we propose targeted methods to reduce stereotypical visual elements in current designs. These methods aim not only to create novel visual representations of characters but also to ensure they remain recognizable to the audience. Second, we seek to mitigate the negative effects associated with stereotypes. We will experimentally test the effectiveness of this method, with success indicated by the character’s representation surpassing existing stereotypes. Finally, we analyze and compare experimental data to evaluate the strengths and weaknesses of the proposed methods, as well as their appropriate applications. In general, this study examines three methods of deconstructing visual stereotypes across three phases, with the data compared and summarized. The methods include: 1) analyzing textual media to deconstruct stereotypes of characters whose origins conflict with current stereotypes; 2) examining multiple media sources and historical records to address stereotypes of characters that have undergone multiple transformations over time; and 3) using ChatGPT-4 to summarize archetypal character stereotypes, then deconstructing these stereotypes to assess whether they are exchanged or retained. This study achieved three primary objectives. First, it proposed specific methods to reduce standardized visual elements, or stereotypes, in current designs. These methods generated new visual representations of the character while ensuring that the character remained recognizable to the audience. Second, the study aimed to mitigate the negative effects associated with stereotypes. Experimental data indicated that the characters used in the experiment broke from existing stereotypes. Given that it takes time for elements of current stereotypes to shift and become widely accepted as new norms, it is theoretically possible to lessen the negative impacts of existing stereotypes by incorporating the character concept into work over time, using the methodology developed in this study. The third objective was to analyze and compare the experimental data to summarize the strengths and weaknesses of the methodologies, as well as the contexts in which they are most appropriately applied. Unlike many studies, this one refrains from directly undermining prejudicial thinking; instead, it proposes multiple methodological approaches. These methodologies are intended as tools for transforming the visual stereotypes of characters. The study employs AI to summarize stereotypes, significantly reducing time demands. It also assesses the success of breaking stereotypes based on the originality and communicative clarity of the work. Overall, this study addresses the limitations and inertia of design thinking, promoting a character-based approach that encourages critical and dialectical thinking among creators. It also provides valuable reference points and inspiration for future research on stereotypes.
 
Keywords—artificial intelligence, stereotypes, character design, image generation


Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).