Dear CITAMS colleagues and friends, I'm so excited to speak to J. Khadijah Abdurahman about the automated prediction of child abuse. Specifically, the AFST algorithm is a system that produces automated recommendations meant to help determine when a child welfare agency should remove children from parents. It's among the more famous examples in recent debates about the societal implications of new computing technologies, and has received notable criticism for the consequences of false positives. In this interactive conversation I expect J. Khadijah Abdurahman will discuss how the continuing push for automation in government services and child welfare agencies has adapted after criticism of AFST, and how newer systems have been linked to outcomes like (for example) pressure for increased surveillance to produce more data that can be used to improve automated predictions of future child abuse. This event is open to the public, please feel free to share it as appropriate. I hope to see you there! Christian ------ TITLE Predicting Prevention: Algorithmic Logics in the Child Welfare System SPEAKER J. Khadijah Abdurahman, The American Assembly and Columbia University in conversation with Christian Sandvig, University of Michigan DATE / TIME / LOCATION Friday, April 23, 2021 4:00 pm - 5:00 pm Eastern Time Online via Zoom HOW TO PARTICIPATE This event is free and open to the public. Click here to join the event: https://umich.zoom.us/j/98631322771 (Note: This Zoom link has changed from the initial announcement of this event.) ABSTRACT In response to the public outcry against the Allegheny Family Screening Tool (AFST), the pilot of predictive analytics in the US child welfare system to predict which child should be removed, state child welfare agencies have justified their use of automated decision systems as a way to “allocate preventive services to at risk communities”. We will move beyond the “garbage in garbage out” analysis of historical inequities baked into the data used by predictive risk models to reflect on what type of infrastructure is built up by algorithms in human services. What kind of resistance is possible to connect the tech labor struggles of engineers tasked with developing these systems and the people on the receiving end of these predictions? BIO J. Khadijah Abdurahman is based at We Be Imagining at Columbia University’s The American Assembly and the INCITE Center. I’m an abolitionist whose research focus is predictive analytics in the child welfare system, the impact of social media and surveillance in Ethiopia. Christian Sandvig is Director of the Center for Ethics, Society, and Computing (ESC) and H. Marshall McLuhan Collegiate Professor of Information, Communication & Media. NOTES Link to this event on the Web: https://esc.umich.edu/event/j-khadijah-abdurahman/ More information about ESC: The Center for Ethics, Society, and Computing: http://esc.umich.edu/ -------- LOOKING AHEAD TO OTHER ESC EVENTS: Dan Greene: The Promise of Access April 21, 2021 @ 1:00 pm - 2:00 pm https://esc.umich.edu/event/dan-greene-the-promise-of-access/ Keith Breckenridge: Recasting The Technologies of the Carceral Empire (part of the series: Behind Walls, Beyond Discipline: Science, Technology & the Carceral State) May 14, 2021 @ 12:00 pm https://esc.umich.edu/event/keith-breckenridge-recasting-the-technologies-of-the-carceral-empire/ -------- ESC is generously supported by the School of Information; the Center for Political Studies at the Institute for Social Research; and the Department of Communication & Media in the College of Literature, Science, and the Arts at the University of Michigan.