Machine Generated Culpability: Inscrutable Machine Evidence in the Criminal Legal Process

Machines used in the criminal legal process increasingly have opaque programming architectures and perform cognitively impenetrable functions. Their inscrutability threatens to undermine the procedures relied on by the criminal legal process to safeguard substantive guarantees such as due process and freedom from arbitrary government action.

This talk, led by BU Law Prof. Ahmed Ghappour, will consider the question of whether and to what extent fact-finders in the criminal legal process should defer to factual outcomes generated by “opaque” algorithms whose form or functionality cannot readily be digested by human-scale observation and reasoning. First, Prof. Ghappour will provide a taxonomy of evidence-producing machines and highlight potential infirmities that may result in improper inferences from machine evidence. Then, he will show how courts struggle to deal with information asymmetries that arise from a lack of access, expertise, and cognitive capacity, producing results that are inconsistent, unpredictable, and at times incoherent. Finally, he will argue that regulation of machine evidence in the criminal legal process should account for the unique characteristics of modern machines, and will make suggestions tailored to these concerns.

Please RSVP to Tyler Gabrielski at tgabs@bu.edu.