AI and Future-Making: Design, Biases, and Human-Plant Interactions

Design researchers and practitioners are turning to generative AI (genAI) to support activities such as ideation and concept development in pursuit of preferred futures. At the same time, genAI is known to have biases, which prompts questions about how these biases might adversely affect design practices. In the domain of sustainable HCI, with its recent trends in human-nature interactions and more-than-human design, the question can be further refined into whether and how genAI biases might perpetuate anthropocentric biases that these practices are increasingly seeking to confront. In the present research, we conducted three workshops, focusing on genAI for human-plant interactions; in the first workshop, we created design fiction concerning human-plant interactions in Southern California in the year 2100, building on the second, the third workshop sought to identify and bring into focus relevant units of analysis. Results included the identification of three kinds of AI biases in plant representations that affect design practices of future-making: species, ecologies, and interactions.