Journal article
Exploring GenAI Image Generation in Engineering: A Thematic Analysis of Ethical and Representational Biases
The International journal of engineering education: Special Issue Artificial Intelligence in Engineering Education – Part I, Vol.41(6), pp.1462-1472
2025
Metrics
13 Record Views
Abstract
Generative artificial intelligence (GenAI) platforms are widely used to create images based on textual inputs. While these tools hold great potential to influence societal perceptions, they also risk perpetuating stereotypes and biases, particularly in fields like engineering, where stereotypical depictions are commonplace. Engineering images often reinforce traditional views of gender, race, and professional roles, raising concerns about whether GenAI tools can produce visuals that are truly inclusive and representative of diverse groups. This study aims to investigate the ability of two GenAI tools, Copilot and Canva, to generate diverse and inclusive representations of engineering teams. By analysing the quality and diversity of images generated using a systematic approach, the research provides insights into the biases embedded within these tools. The quality of the generated images had several flaws, including multiple missing or additional limbs, facial features, or fingers. This study revealed the flawed and generic images that GenAI tools can generate when asking to generate a simple engineering team. Racial, gender, and age-basedstereotypes of engineers were a constant theme throughout the images. There was limited diversity and inclusion of Indigenous peoples, hair colour and length, and body shape and size. The findings are intended to set a baseline for future evaluations and improvements in GenAI platforms.
Details
- Title
- Exploring GenAI Image Generation in Engineering: A Thematic Analysis of Ethical and Representational Biases
- Creators
- Zachery Quince - Southern Cross UniversityKathy Petkoff - Monash UniversityAnna Lidfors Lindqvist - University of Technology Sydney (UTS)Emily Faulconer - Monash UniversityWinn Chow - University of MelbourneSasha Nikolic - University of Wollongong
- Publication Details
- The International journal of engineering education: Special Issue Artificial Intelligence in Engineering Education – Part I, Vol.41(6), pp.1462-1472
- Publisher
- IJEE
- Identifiers
- 991013329886102368
- Academic Unit
- Centre for Teaching and Learning; Faculty of Science and Engineering
- Language
- English
- Resource Type
- Journal article