Recap (AI, Law, and Otter Things #21)
Happy New Year, dear readers! This first issue of 2022 will be a bit more personal than usual, as I write about what I did and what I liked in 2021. At the end of this newsletter, I will share some book recommendations, so you might want to skip ahead to that or come back for the next issue.
One of the first issues of this newsletter provided a recap of my first academic year as a doctoral researcher. Back then, I had set some goals for myself:
Explore new directions of research instead of staying in my comfort zone;
Write more stuff by myself;
Find a regular rhythm for my work; and
Reaching a broader audience.
Since then, I have been pretty happy with my progress towards the first three goals. In the second semester of this year, I completed fewer texts than I had in the previous months, giving me the time to explore new topics and research approaches. Exploration can be a woefully inefficient process, and indeed it feels a bit weird not to be collecting many low-hanging fruits from my work. But, in the last few months, I took the time to be a bit speculative with my thesis project, and now I believe I have something interesting to say. It also gave me some ideas for side projects, which I hope to share here in due time.
Given my emphasis on exploration and longer-term projects, this newsletter turned out to be a good outlet for my need for quick wins. A few times over the last few months, I found myself writing a few paragraphs over a topic that caught my attention rather than allowing myself to try and develop it into another side project. Doing so not only helped me focus on my main lines of research but also led to some exciting discussions with readers. And, more importantly, the newsletter made sure I was writing something every week (or fortnight) rather than simply rushing text to meet deadlines. So, it has helped me to cultivate some healthy writing habits.
The road ahead
I am, however, less content with my progress in reaching a broader audience. On the one hand, it is only expected, as I haven't been to many events, and my English-language outputs from 2021 are still unpublished (more on that over the following weeks, hopefully). On the other hand, I feel like I could've done more on this front, especially when it comes to meeting people working on similar topics. So, I plan to reach out more actively to my peers in 2022 (and please feel free to drop me a line if you'd like to chat!).
Another missed goal that I'll resume this year is that of refreshing my knowledge of math and the technical side of AI. While my limited professional experience suggests that many critical applications rely on technologies far from the bleeding edge of technological development, I feel that getting closer to the technical side of things can be useful for the kind of arguments I deal in. And it can be fun, especially if you are not doing it professionally.
I have also decided to do something I've been postponing for more than a decade: actually learning some French. Now that I am more comfortable with my Italian, I am running out of excuses for not doing so, especially as this language might be helpful for my career prospects.
Last but not least, some colleagues and I are launching The Digital Constitutionalist blog, a space for discussing the implications of digitalisation for fundamental rights and the limitation of powers, both those exercised by states and by online platforms. We will also have a section for essays and original science fiction works exploring how sci-fi influences our thoughts about those themes. I will write more on the blog when we get closer to its launch date, but I believe it will be of interest to many of the readers of this newsletter.
What I did in 2021
When it comes to publications and work presentations, most of what I did in the past year was done or at a very advanced stage by the end of the first semester. For my own convenience, I will recap things here, adding the novel content from the second semester and distinguishing English- and Portuguese- language outputs. To the extent possible, I have tried to provide links to the relevant material, but please contact me if you would like to read any of these papers.
My main research output was my May Paper, titled ‘A techno-legal approach to the life cycle of artificial intelligence systems’, which I defended as part of the first-year requirements in the EUI PhD in Law. There, I sustain that life cycle models from software engineering can be used to empower legal assessments of AI systems by facilitating communication with technical stakeholders and the diagnosis of the ideal points for legal intervention. Part of this paper was further developed in a presentation for ETHICOMP 2021, titled ‘Some Remarks on the Ethical Relevance of Leaky Abstractions’, where I argue that algorithm-centric accounts of AI overlook important aspects of these technologies.
Beyond my thesis, I also had the opportunity to present some work developed in projects I have joined. My involvement in the Un-owned personal data project has led me to participate in a few events, in particular a workshop hosted by the EUI in November. I also had the opportunity to develop the paper ‘A Robot and a Moderator Walk into a Bar: The Use of AI in Online Moderation of Humoristic Content’ with my wife, Renata Shimbo. Recently, she has been deploying her background in humour theory to study the challenges of moderating humouristic content in online platforms, and our paper, presented at a TMC Asser workshop in September, explores how these challenges impact automated moderation practices.
My long-standing collaboration with Juliano Maranhão has led to various publications this year. Last year, we had worked together on the regulation of healthcare automation, leading to the English-language article ‘Voice-Based Diagnosis of Covid-19: Ethical and Legal Challenges’, published on a special issue of IDPL. This line of research also led to a book chapter in Portuguese, ‘Inteligência Artificial no Setor de Saúde: Ética e Proteção de Dados’, published on an edited volume on data protection law and the healthcare sector.
Maranhão and I have also done some work in establishing interdisciplinary communication between Brazilian legal scholarship and computer science. This interdisciplinary effort had three main outputs, all in Portuguese. Back in the first semester, we gave a talk to the São Paulo School of Judges on the use of artificial intelligence in notarial and registry tasks. We also wrote a article with Juliana Abrusio, ‘Inteligência artificial aplicada ao direito e o direito da inteligência artificial’, which discusses the technical challenges involved in implementing regulation and legal AI tools that might be used to address them. This article was published in Suprema, the academic journal maintained by the Brazilian Supreme Court. Finally, Fabio Gagliardi Cozman joined Maranhão and me for a book chapter on explainable AI, titled ‘Concepções de Explicação e Do Direito à Explicação de Decisões Automatizadas’, where we survey current explanation techniques and argue that explanations in the legal domain should be conceptualised as contestability-enabling tools.
Finally, I also was involved in some teaching-related activities this year. I taught the doctoral seminar ‘AI in Practice: Technical and Legal Perspectives’ with a few colleagues at the EUI. I also acted as an administrative assistant to the Law and Logic and AI and Law summer schools, and I got involved in the Technological Change and Society research cluster at the EUI, roles I intend to repeat this year.
The past year was a bit too intense, especially in the first semester, but all in all, I am quite satisfied with what I did in 2021. Especially because this narrative only presents the successes, hiding the various rejections, failures, and false starts. I like the idea of building a CV of Failures, and I plan to come back to this topic in a future issue. But, for now, I will allow myself to bask in the victories.
Readings from 2021
At first, I planned to write a few "Top 5 lists" here: my favourite reads from 2021, TV shows, etc. But I ended up writing too much in my recap, which means I will narrow down my focus to a few reading tips that I enjoyed over the past year.
As a non-EU scholar working in a field where EU law plays an important role, I've been trying to become more familiar with this broad speciality of law. For newcomers like me, Schütze's European Constitutional Law might be an interesting follow-up to the introductory textbooks on EU law. In an attempt to get a broader knowledge about the Union's institutional arrangements, I asked Twitter for some recommended readings, and it did not disappoint:
If you are interested in the development of EU institutions, follow the replies to my question and you'll find some excellent materials.
Another gap in my formation that I've been trying to address is that I am not really familiar with qualitative social research. My software engineering courses covered some approaches, such as ethnographies and interviews, but with a very different emphasis than what I saw in empirical legal scholars. So, I've tried to learn more about methodology, and one book I found particularly useful is Halliday and Schmidt's Conducting Law and Society Research. The authors interviewed various law and society scholars about their approaches, and in doing so, they present lots of practical advice that is often buried in more formal methodology books.
When it comes to my focus on AI and regulation, there are a few materials that might be interesting for readers of this newsletter. Your Computer Is on Fire is an excellent collection of critical essays about technological infrastructure, which bring to the fore how online platforms, AI systems, and other forms of digital technologies are contingent upon human labour and knowledge and situate them in history. In particular, I have enjoyed how they present and frame the brittleness of the technological underpinnings of our information society. This infrastructural role of information technologies has profound legal implications, and I found Julie Cohen's treatment of the techno-legal interface in Between Truth and Power quite insightful, as it looks at information and communication technologies in their socio-technical contexts, as opposed to ascribing them infinite malleability in the face of regulation or, in the opposite direction, surrendering to technologists' often hyped narratives of what technologies such as AI can achieve. Readers might also be interested in Matthijs M Maas's discussion of AI and socio-technical change and Till Straube's analysis of the limits of the black box metaphor.
Moving on to fictional readings, I really enjoyed Philosophy through Science Fiction Stories as a philosophical exercise and as a source of good short stories. Ken Liu's short story, which deals with the use of brain augmentation technology by lawyers, provides an excellent example of how sci-fi may be used to think about questions relating to technology law. A Memory Called Empire blew me away with its worldbuilding, especially as it draws from under-tapped historical sources, and it captures well the sensation of being immersed in a new culture that you have studied from afar. And The Ministry for the Future, for all its shallow characters, still provides us with a climate story that avoids the trap of monocausal explanations for complex phenomena.
This is it for today, and now I leave you with a sleepy otter: