Critical Analysis of Law Workshop Series
presents
Mireille Hildebrandt
Vrije Universiteit (Brussels)
Privacy as the protection of the incomputable self
Issues of data protection law in a data-driven environment
Tuesday February 13, 2018
12.30 – 2.00 pm
Solarium (room FA2), Falconer Hall
84 Queen’s Park
Not everything that can be counted counts, and not everything that counts can be counted. Based on this crucial insight I will propose a new dimension to our understanding of privacy, as the protection of the incomputable self. I will argue that in the era of big data analytics, we need an understanding of privacy that is capable of protecting what is uncountable, incalculable or incomputable about individual persons without, however, rejecting the use of machine learning. I will explain that one way of protecting what cannot be counted, datafied or inferenced is to require ‘agonistic machine learning’, i.e. demanding that companies or governments that base decisions on machine learning must enable the development of alternative ways of datafying and modelling the same event, person or action. This should ward off monopolistic claims about the ‘true’ or the ‘real’ representation of human beings, their actions and the rest of the universe in terms of data and their inferences (cf. McQuillan’s machinic neo-platonism). Building on Mouffe’s concept of agonism in the context of democratic theory and Rip’s concept of agonism in the context of constructive technology assessment, I develop a notion of agonism that fits the upcoming legal obligations in the GDPR, notably to (1) conduct a data protection impact assessment and to (2) implement data protection by design. I will link this with the principle of purpose binding that is core to the GDPR. Though some authors claim that the legal principle of purpose binding as articulated in the GDPR is out of tune with the reality of big data analytics, I will align with Van der Lei and Cabitza who explain that, on the contrary, techniques such as machine learning assume purpose specification to make sense of big data (= the first law of informatics). Indeed, as Mitchell explains when providing a definition of machine learning, the latter requires a machine-readable task and a machine-readable performance metric that both assume the specification of a purpose. Finally, I will argue that the combination of purpose specification and agonistic machine learning will enhance both the methodological integrity of machine learning and our capability to protect what cannot be counted, computed or inferenced.
Mireille Hildebrandt is a lawyer and a philosopher. She is Research Professor of ‘Interfacing Law and Technology’, appointed by the Research Council of the Vrije Universiteit Brussel at the VUB Faculty of Law and Criminology, and part-time Full Professor at the Science Faculty, department of Computing Science, at Radboud University, Nijmegen. Her work is focused on the nexus of postive law regarding cybercrime, data protection and human rights, legal theory, and philosophy of technology. Hildebrandt has published extensively on these subjects (4 books, 21 edited volumes, 104 chapters and articles in scientific publications). She is teaches ‘Law in Cyberspace’ to master students of computer science and currently teaches the intensive course on ‘Data Driven Law’ as Distinguished Visiting Professor at University of Toronto Law School.
A light lunch will be provided.
For more workshop information, please contact Nadia Gulezko at n.gulezko@utoronto.ca.
