Automated decision tools are increasingly significant technology in the economy and this trend has been particularly true in the welfare system where automated decision support tools are used to determine eligibility for unemployment insurance, Medicaid or allocate subsidized housing. With rollout of these tools, there has been a rise in concerns about potential for bias concerns these may disproportionately hurt disadvantaged minorities. While much of the literature has focused on specific algorithms, data inputs and outcome variables, the institutional and strategic context in which these tools are developed and sold is also likely to play a role in the eventual impact of these tools. In this paper, I highlight the role of intellectual property ownership of these tools and the impact of the IP rights on use of decision support tools by Child Protective Services agencies in the United States.
