In the second video in COHUBICOL’s ‘interview with the author’ series, our researcher Emilie van den Hoven spoke to Sylvie Delacroix (@SylvieDelacroix) in the context of our 2020 Philosophers’ Seminar, held earlier in December. For the seminar we convened scholars from philosophy, computer science and law to foster a cross-disciplinary exploration of some of the most pertinent questions in relation to the use of AI in law. The focus of this year’s session was on the ‘interpretability problem’ in machine learning.
In the interview Delacroix talks about her paper ‘Diachronic interpretability and automated systems’ and discusses how our normative agency can atrophy if the automated systems used in morally-loaded contexts are not designed in such a way as to preserve our interpretive capabilities and our capability for normative decision-making. The machine-enabled ‘normative holidays’ that can consequently result can have devastating impacts, and ought to be avoided.
Please enjoy the second COHUBICOL video of our ‘interview with the author’-series!