Troubling taxonomies in GenAI evaluation

To evaluate the societal impacts of GenAI requires a model of how social harms emerge from interactions between GenAI, people, and societal structures. Yet a model is rarely explicitly defined in societal impact evaluations, or in the taxonomies of societal impacts that support them. In this provocation we argue that societal impacts should be conceptualised as application- and context-specific, incommensurable, and shaped by questions of social power. Doing so leads us to conclude that societal impact evaluations using existing taxonmies are inherently limited, in terms of their potential to reveal how GenAI systems may interact when deployed in specific social contexts. We therefore propose a governance-first approach to managing societal harms attended by GenAI technologies.