Governing AI to Advance Shared Prosperity

Abstract

This chapter describes a governance approach to promoting AI research and development that creates jobs and advances shared prosperity. Concerns over the labor-saving focus of AI advancement are shared by a growing number of economists, technologists, and policymakers around the world. They warn about the risk of AI entrenching poverty and inequality globally. Yet, translating those concerns into proactive governance interventions that would steer AI away from generating excessive levels of automation remains difficult and largely unattempted. Key causes of this difficulty arise from two types of sources: (1) insufficiently deep understanding of the full composition of factors giving AI R&D its present emphasis on labor-saving applications; and (2) lack of tools and processes that would enable AI practitioners and policymakers to anticipate and assess the impact of AI technologies on employment, wages and job quality. This chapter argues that addressing (2) will require creating worker-participatory means of differentiating between genuinely worker-benefiting AI and worker-displacing or worker-exploiting AI. To contribute to tackling (1), this chapter reviews AI practitioners’ motivations and constraints, such as relevant laws, market incentives, as well as less tangible but still highly influential constraining and motivating factors, including explicit and implicit norms in the AI field, visions of future societal order popular among the field’s members and ways that AI practitioners define goals worth pursuing and measure success. I highlight how each of these factors contributes meaningfully to giving AI advancement its excessive labor-saving emphasis and describe opportunities for governance interventions that could correct that over emphasis.