Tinker, Tailor, Configure, Customize: The Articulation Work of Contextualizing an AI Fairness Checklist

DOI

Many responsible AI resources, such as toolkits, playbooks, and checklists, have been developed to support AI practitioners in identifying, assessing, and mitigating potential fairness-related harms. These resources are often designed to be general-purpose, to be applicable to a variety of use cases and contexts; however, this often leads to de-contextualization, where they may lack relevance or specificity for AI practitioners developing particular applications for specific contexts. To understand how AI practitioners may re-contextualize one such responsible AI resource, an AI fairness checklist, for their particular contexts and AI applications, we conducted semi-structured interviews with 13 AI practitioners from seven organizations. We identify how contextualizing the fairness checklist produces new forms of work for AI practitioners and other stakeholders, as well as opens up new sites for negotiation and contestation via decisions about how to adapt the fairness checklist. As part of this, we identify how this contextualization process may help AI practitioners develop shared language about AI fairness, and we identify tensions in ownership over this process that suggest larger issues for accountability in the work of responsible AI.