$5k USD grant from Google Research for interview incentives + transcription costs. Project description below:
Industry and university-based RAI researchers have expended significant energy developing taxonomies that attempt to map the deployment of AI-related sociotechnical systems to particular potential (negative) outcomes or risks. There is anecdotal evidence that these taxonomies are beginning to be used, particularly in industry-based product teams and also in government agencies, to support AI practitioners to reflect on and mitigate societal risks associated with AI deployment. However, as yet, there has been no systematic analysis of the effectiveness of RAI taxonomies. To address this, we intend to undertake an interview-based study, which will recruit two cohorts of participants: first, RAI taxonomy developers in industry and academia will be interviewed; second, RAI taxonomy users in industry and government will be interviewed. Together, these interviews will enable the first qualitative analysis of the role RAI taxonomies are currently playing in the RAI ecosystem.