Deconstructing the Algorithmic Gaze: A Visual Analysis with Midjourney

This workshop turns you into a researcher of AI bias. Using Midjourney and other image generation tools, we will systematically prompt these systems to produce images of the Middle East — cities, people, historical events, everyday life — and then critically analyze the results. What visual vocabularies do these systems default to? Where do the tropes come from? What is absent?

We will work through a structured protocol: generating images with identical prompts across different tools, documenting patterns of representation, comparing AI-generated images to real photographs, and tracing the visual genealogy of specific tropes back to Orientalist painting and colonial photography. This is not an art workshop — it is an exercise in excavating the assumptions embedded in training data.

The workshop connects directly to Crawford’s “Excavating AI” and Kotliar’s “Data Orientalism,” grounding those theoretical frameworks in hands-on experimentation. You will leave with a portfolio of generated images annotated with your critical analysis — material that can inform your final research project and, more broadly, your ability to see the ideological work that AI image systems perform.