Some sparse notes (AI, Law, and Otter Things)
Dear reader, welcome to the latest issue of my newsletter! Today's issue is something of a filler episode: I kinda forgot that the newsletter is now released on Wednesdays when doing my weekly schedule. Since this week I am also participating in a workshop on AI, Business and Human Rights, I ended up with little time to finish a proper newsletter issue. So I will present some interesting stuff I've read on tech and law, in topics that might be of interest to my average reader.
This newsletter's previous issue finished with a call for references on the subject of future-proof regulation. Since then, I have learned about a volume edited by Sofia Ranchordás and Yaniv Roznai, Time, Law, and Change: an Interdisciplinary Study (2020). Its sixteen chapters deal with various issues regarding temporalities in the law, such as sunset clauses and experimental regulation, and two of the chapters are particularly relevant for law and technology. The first one, by Ranchordás and Mattis van't Schip, analyses how experimental legislation and impact assessments can be used to ensure regulatory robustness. It does not provide a full conceptual analysis of "future-proofing", but it addresses some important criticisms of future-proof approaches and has exciting proposals regarding what we can actually do when it comes to regulation.
A second chapter that is relevant for this discussion is authored by Lyria Bennett Moses and Monika Zalnieriute. There, the authors dismantle the myth that technology is always lagging behind regulation. Drawing from legal and science and technology studies (STS) resources, the authors paint a fuzzier portrait: the law is indeed slower to change than technology, software often finds itself in the role of catching up with legal change, too, as a result of both technical path dependency and organisational factors. Even though the paper stumbles into one of my pet peeves (I am not a fan of the term "data privacy"), it is a valuable call to interdisciplinary action and a richer understanding of technology as a legal object.
These conceptual debates, in turn, are relevant for understanding the broad array of legislation currently being proposed to regulate various aspects of digital society. Verfassungsblog currently has a blog debate dealing with specific proposals of European Union regulations: the Digital Markets Act and the Digital Services Act. All the contributions to this debate deal, in one form or another, with the issue of private power, and a few of them are particularly insightful. Giovanni De Gregorio and Oreste Pollicino present how these regulatory approaches reflect a move towards a strategy of digital constitutionalism geared towards limiting private power and protecting fundamental rights while trying to put reasonable limits on their horizontal effect. Naomi Appelman, João Pedro Quintais, and Ronan Fahy point out that Article 12 DSA, despite its potential to require attention to fundamental rights in platforms Terms and Conditions, is too vague in linguistic terms to provide any meaningful guarantees. Also on the subject of private enforcement of fundamental rights, Peter Picht discusses some points that must be addressed to ensure that private actors play their intended roles in the enforcement of a system that might be too complex to rely solely on active public oversight.
After these recommendations, I would like to conclude with a shameless plug for my lusophone readers. A new book on AI, society, economy and the state is available for preorder, and it includes a chapter I wrote with Juliano Maranhão and Fabio Cozman on explainable AI (XAI). There, we present the state of the art on XAI technologies, relating them to the various views on the goal of explainability and arguing that, for legal purposes, what matters is that we approach explanation as a task geared towards enabling contestability of decisions involving AI. This chapter is currently available only in Portuguese, but I hope to have something on the topic in English at some point in the future.
Well, that is it for today. See you next week, hopefully with a full-length issue!