Hello, dear reader, and welcome to the 46th issue of AI, Law, and Otter Things!1 Since the last post, I have mostly focused on reworking a paper that got a “Major revisions” result from peer review. I am not entirely happy with the outcome of my revisions, especially after spending more time on them than I should have, but I don’t think I can make things much better than the current version of the manuscript. So, I will use today’s newsletter to write a bit about work processes.
After that, this newsletter will follow the usual format. I will share some readings that I found interesting and wrap up with the usual otter picture. But, before that, I will share with you a photo of Winnie, who has been absent from the latest issues because she is not very fond of being photographed:
My writing process
As I mentioned in the previous issue, I’m not a big fan of approaches to work that require intensive focus, such as deep work or even the classic Pomodoro method. Some friends have obtained excellent results with methods like these, but I’ve always struggled to comply with the kind of strict time management they require. And, even when I could bring myself to do so, my outputs were quite uninspired, to say the least. So, it took me a long—and still ongoing—process to find out what works for me.
My personal workflow has to address a few shortcomings of mine. The first one is that I am not a particularly deep thinker, who can grab hold of a subject and leave no stone unturned; instead, my contributions, if any, tend to involve looking at things from new angles or identifying tensions in existing approaches.2 The second one is that many of my best insights come at random moments, but actually developing such insights into something meaningful will always require actual work rather than relying on a spark.3 Finally, I really enjoy the communication that writing enables, but I’m not particularly fond of the writing process itself.4 I like to think that none of these is, in itself, an obstacle to an academic career,5 but overlooking their potential adverse effects would be a foolish move.
So, I try and follow some rules of thumb whenever I’m writing. Since I am very much a morning person, I try to schedule my writing tasks in the morning,6 and I try not to write for more than two hours at a time. In fact, most of my writing tends to take even smaller chunks of time: since my ideas appear at random moments, I try to write them down as soon as possible,7 and my larger writing blocks are mostly dedicated to making sense of these notes and shaping them into cohesive text. This approach also takes advantage of the fact that I am not particularly attached to my own writing, so I find it easy to substantially change what I wrote in the past and even discard large chunks of it.
I make no claim that everybody, or even most people, should follow similar heuristics. Scholars who follow depth-first approaches to research might benefit from more focused approaches, and I am well aware that some of these constraints to work are not always feasible. For example, writing in the mornings was not really an option when I was working full-time and writing my Master’s thesis, and deadlines often push me to write for more than two hours at a time (though I always take a long time to recover after sustained efforts). If anything here is generalizable, it is the idea that you should know your constraints and limitations and try to organize your workflow around them as much as possible.
For that purpose, I stole picked up bits and pieces from various people who write on methods and writing. Raul Pacheco-Vega has posted many amazing resources on how to organize your research, ranging from the most practical issues on writing to the more strategic questions on how to prioritize things and organize your time. Eviatar Zerubavel’s The Clockwork Muse is a handy guide to time management, particularly useful for those of us that have had to juggle research, family life, and external work. From Niklas Luhmann, I got the idea of frequently writing small notes and using them as tools for thought, even though I found the Zettelkasten method too formal for my needs. And, more recently, William Germano’s On Revision really upped my revision game.8 These and other sources can help you find interesting ideas to test, and even if you don’t stick with a particular method or approach, it might give you some ideas on how to adapt your workflow to your needs.
Events and calls
Recently, my supervisor (Deirdre Curtin) published a volume she edited with Mariavittoria Catanzariti on Data at the Boundaries of European Law. My friends at the EUI Digital Public Sphere workgroup will host a launch event on 3 May, in which the editors will present the book’s contributions and discuss them with Madalina Busuioc, Herwig Hofmann, Païvi Leino-Sandberg, and Gianclaudio Malgieri.
The deadline for extended abstracts for CRCL’s conference on “Computational ‘law’ at the edge” ends tomorrow.
For those of you who are more sociologically-minded than me, Socius has an open call on “A Sociology of Artificial Intelligence: Inequalities, Power, and Data Justice”. They are looking for 300-word abstracts until 1 July.
On 27 and 28 April, the Solvay Brussels School of Economics and Management will host the conference “A 'Brussels Effect' for EU digital governance”. On the 27th, Anu Bradford will kick off the works with a keynote, while the 28th will be dedicated to presentations by early career scholars and a final panel on “Beyond the Brussels Effect”. I will be there, presenting a paper on the AI Act’s external effects that I’m developing with Anca Radu. Registration is required but free, and if any readers are in Brussels on the 26th or the 27th, please feel free to get in touch.
Reading recommendations
Before suggesting some interesting work by others, I will seize the opportunity to point you to my recent blog post for the Robotics and AI Law Society (RAILS). In that post, I draw from Michael Guihot’s work on technology law to suggest that AI regulation is better understood not as a unified field of study but as a loose coalition of scholars with overlapping interests. Instead of pursuing an internally consistent framework to talk about AI in the law, perhaps we are better served by valuing this diversity and finding ways to learn from one another.
Fabian Beigang, ‘Reconciling Algorithmic Fairness Criteria’ [2023] Philosophy & Public Affairs Early access.
Discussions of algorithmic fairness have long been plagued by an impossibility result that suggested that algorithms could not simultaneously attend to some of the established definitions of fairness. Recently, some authors have proposed alternative metrics and potential trade-offs, while others have argued for a shift towards substantive fairness. While I am partial to the latter camp, Beigang’s paper suggests that a causal perspective may be used to reconcile the seemingly incompatible criteria for algorithmic fairness. So, it seems formal accounts of fairness might not be as dead as initially thought.
Cristina Cocito and Paul De Hert, ‘The Use of Declarations by the European Commission: “Careful with That Axe, Eugene”.’ (The Digital Constitutionalist, 14 March 2023)
As this issue’s title suggests, I am in a bit of a Pink Floyd mood these days. This mood prompted me to take the time to read the post linked above, which I now recommend to you as an interesting analysis of the practical effects of Commission soft law and how their misuse can create problems for the EU approaches to digital law.
Nicolas Kayser-Bril, ‘The ideology behind publishing Twitter's source code’, Automated Society.
AlgorithmWatch’s newsletter always has some interesting content, and this time they included a useful discussion on why source code disclosure is not that much of a form of disclosure.
Paul Ohm, ‘The Argument against Technology-Neutral Surveillance Laws’ (2010) 88 Texas Law Review 1685.
Paul Ohm is one of these technology law scholars whose work I keep revisiting. I tend to disagree with him quite a bit, but he always takes the implications of digital technologies for legal issues seriously and does not shy away from the technical bits. In this article, Ohm responds to the dogma of technology neutrality in regulation and suggests that the risks of technology-specific legislation are mitigated by the strong background rules that discipline surveillance and that their adoption has practical and political advantages that one would miss by focusing on neutrality as desirable in itself.
Siddharth Peter de Souza and Lisa Hahn, The Socio-Legal Lab: An Experiential Approach to Research on Law in Action (Open Press Tilburg University 2022).
Modern research on the law is increasingly reliant on methods and questions drawn from the social sciences. While I think that legal academia sometimes goes too far in the direction of socio-legal studies, it is undeniable that such approaches have contributed to our understanding of how legal phenomena play out in society. But lawyers, especially those of us trained in Civil Law jurisdictions, are not always educated in the methods of the social sciences. This book can be a useful tool to understand what it means to study law and action, helping both those scholars interested in deploying socio-legal methods and those seeking to understand their limits.9
Bonus track for Portuguese speakers: Os Últimos Samurais, TAB UOL.
Early in the 20th century, Brazil received a huge influx of Japanese migrants, who established themselves mostly in the Southeastern states. To this day, Brazil hosts a large share of the Japanese diaspora, with more than 1.5 million Brazilian of Japanese ancestry (including my in-laws!) living in the city of São Paulo alone. The news feature linked above focuses on a harrowing passage of that story: after the end of the Second World War, the Shindo Renmei organization refused to accept the Japanese surrender, running a campaign of fake news and attacks against the “dirty-hearted people” that spread the news about the actual outcome of the war. The feature presents the story of the movement, from its origins in the 1920s to its terror campaign and subsequent repression, and how the whole affair is seen in the present day.
And now, the otters
See you next time! And, if you haven’t done so yet, please subscribe to receive future updates via email.
Or, as I say half the time, Otter Stuff. I wonder if I should change the newsletter’s title.
In the hedgehog-fox spectrum, I am solidly on the side of foxes.
As a wise man once put it, “Ninety percent of most magic merely consists of knowing one extra fact.”
Not just of writing in English, despite the additional work involved in using a foreign language. If anything, things get even more awkward when I have to write about work (or quite a few of my personal interests) in Portuguese.
Though they do create restrictions on the kind of work one might do. See, e.g., Liam Kofi Bright’s argument that syncretic approaches to philosophy are unlikely to outlast their particular contexts.
This approach also has the advantage of allowing me to write down some words before the kinds of urgent stuff or breaking news that might derail the rest of the day. I imagine that night owls have a similar advantage, as they work while people like me sleep.
Usually, I take notes on Notion or Signal’s “Note to Self” function. Paper-based approaches such as bullet journals do not work for me, as my handwriting is very slow and I enjoy the search functionalities available for digital notes.
Long-time readers of this newsletter might be a bit suspicious, given the Grauniad-like issues that afflict previous issues. However, part of my idea in creating a newsletter was precisely to find a space in which I can write without rewriting everything a thousand times.
Not that these categories are disjunct, of course.