Short abstract:

The paper discusses a provocative foray into deliberative research with Chat-GPT. It asks why the media of deliberative practice so often seem to evade sociotechnical critique.

Long abstract:

Deliberative research methodologies have long been instrumentalised for producing public responses to technoscientific uncertainty. Contingencies and machinations involved in generating informed deliberation have now been widely scrutinised, highlighting how expert curation of topical resources can frame public issues and demarcate the scope of legitimate contestation. Still, relatively limited attention has been given to the roles media-technologies play in the practice of small-group deliberation. This paper will ask why the media of deliberative practice so often seem to evade sociotechnical critique. In pursuing this question, I’ll discuss a methodological provocation that involved deploying Chat-GPT in deliberative research addressing technologies for carbon dioxide removal. By introducing the highly sensationalised “AI” device, we sought to explore whether participant generated texts would impact focused discussion and, specifically, how uncertainties around machine-learning technologies might enter into appraisals of carbon removal futures. The paper will discuss some of the different configurations of expertise, media and uncertainty generated in 4 UK-based workshops. It will reflect on some distinctive problems that generative media-technologies may pose for deliberative research and how they might be appropriated for experiments in participation.