Send message to Author

Short abstract

This contribution explores the potential and pitfalls of AI images as an anticipatory and speculative method for urban transition, examining how image-generative AI intervenes in material urban processes and asking how these tools can offer a democratic infrastructure for shaping urban futures.

Long abstract

AI-driven image generation offers a new and tempting anticipatory method for envisioning urban futures, providing seemingly perfect, friction-free visions of sustainable cities while ignoring uncertainty and complexity. In my contribution, I will discuss the potential and pitfalls of AI images as anticipatory and speculative method for urban transition, examining how image-generative AI intervenes in material urban processes. Based on preliminary fieldwork, I analyze three domains of use: (1) professional practices among architects and urban designers, particularly in scenario planning and speculative design (e.g., Zaha Hadid); (2) non-professional applications in participatory visioning (e.g., UrbanistAI / Kiezlabor Berlin); and (3) political uses in foresight (e.g., Jared Kushner on Gaza City). While the widespread adoption highlights the coercive power of images and the influence of generative AI, it also raises critical concerns: energy consumption, right-wing aesthetics, western-centrism, copyright, or big-tech monopolization, to name just a few.

Ultimately, the question is how AI images can help imagine a future worth living in and creating. Drawing on Comi and Whyte’s (2018) framework, I investigate whether AI-generated images can serve as epistemic artifacts for learning and transformation—rather than instrumental tools or teleological ends. Applying an STS perspective, I examine the assumptions users bring to these tools and those inscribed in their technological structure. The central question I aim to address remains: whose hopes and vulnerabilities are accounted for when envisioning urban futures through AI—and how can we ensure these tools, within their structural constraints, contribute to a fair and democratic infrastructure for shaping material urban processes?