Amit Haim

I study the use of algorithmic decision-making tools, based on machine learning techniques and predictive analytics, in government settings — particularly in child welfare and protection services. Algorithms are becoming more prevalent in government settings in general, but pose even greater challenges when implemented in welfare contexts, especially in situations as sensitive as child protection. My research examines, through empirical inquiry as well as legal analysis, how algorithms are implemented in agencies in practice, how they shape workers’ roles and discretion, and what are and should be the governance mechanisms. Welfare bureaucracies are undergoing technological transformations, and my research is intended to inform these changes and ensure that the problems that occur are addressed, while opportunities for improvement are not missed.