Pixels and Pedagogy: Examining Science Education Imagery by Generative Artificial Intelligence

Abstract

The proliferation of generative artificial intelligence (GenAI) means we are witnessing transformative change in education. While GenAI offers exciting possibilities for personalised learning and innovative teaching methodologies, its potential for reinforcing biases and perpetuating stereotypes poses ethical and pedagogical concerns. This article aims to critically examine the images produced by the integration of DALL-E 3 and ChatGPT, focusing on representations of science classrooms and educators. Applying a capital lens, we analyse how these images portray forms of culture (embodied, objectified and institutionalised) and explore if these depictions align with, or contest, stereotypical representations of science education. The science classroom imagery showcased a variety of settings, from what the GenAI described as vintage to contemporary. Our findings reveal the presence of stereotypical elements associated with science educators, including white-lab coats, goggles and beakers. While the images often align with stereotypical views, they also introduce elements of diversity. This article highlights the importance for ongoing vigilance about issues of equity, representation, bias and transparency in GenAI artefacts. This study contributes to broader discourses about the impact of GenAI in reinforcing or dismantling stereotypes associated with science education.