More or Better Climate Stories? On the (Necessary) Limits of GenAI in the Environmental Humanities

Talk with Dr. Bryan Yazell (University of Southern Denmark).

The environmental humanities, otherwise spanning a diverse field of methods and perspectives, is united by a consensus regarding the social function of narrative. In simple terms, this consensus affirms that stories shape the way we interact with the world and our expectations for the future. At the same time, popular narratives in the West tend to suffer from several imaginative failures that, in turn, compromise our ability to visualize—and therefore respond to—climate change. In response to what many describe as the cultural crisis driving anthropogenic climate change, prominent strands of research in the environmental humanities are committed to uncovering more stories that can fill the imaginative gap in dominant narrative forms, whose embedded anthropocentricism is a primary driver of climate change. 

This latter trend is at the forefront of this talk, which appraises recent scholarship that attempts to leverage generative artificial intelligence (genAI) models such as Chat-GPT to supplement what are, by now, the familiar shortcomings of climate fiction and related Anthropocene narratives. In view of the imaginative problem that climate change poses for fiction, genAI can appear as attractive tools for generating new, or at least defamiliarized, stories. The dramatic environmental costs to genAI creation, now a familiar refrain in critical AI studies, immediately appears as significant downside to this approach. Expanding on this critique, this essay adapts work from AI scholars such as Emily Bender and Kate Crawford to scrutinize the impulse in a strand of environmental humanities research that values the generation of more climate-related stories as valuable in and of itself. This thinking finds an alarming parallel in tech firms, who view the collection of more data—extracted by nearly any means necessary from social media profiles, online message boards, and government databases—as both immensely valuable and necessary for solving any biases or limitations in their genAI models. While the field of environmental humanities has by no means collectively embraced genAI, it is nonetheless crucial to first recognize and reject this more-is-better teleology during a time when researchers and universities face considerable pressure to integrate genAI to make their work more legible to administrators and funding agencies.

Bio

Bryan Yazell (PhD, University of California, Davis, USA) is an Associate Professor of English and American Studies at the University of Southern Denmark and a fellow at the Danish Institute for Advanced Study. He was previously a postdoctoral researcher at the “Uses of Literature” Niels Bohr Professorship Center (2017-2019), where he investigated the intersection between literary fiction and social policy on subjects including migration and social welfare. Research from this project appears in his book, The American Vagrant in Literature: Race, Work and Welfare (2023), which considers authors such as John Steinbeck, Mark Twain, and George Orwell.

Another major focus of his work involves the study of climate change and speculative fiction, especially from an interdisciplinary perspective. More recent scholarship in this area, which examines topics such as climate-induced migration, has developed alongside collaborations with partners in the social sciences, biology, health, and citizen science.

Contact

Contact person for more information: Dr. Martyn Bone (bone@hum.ku.dk)