The Brussels Effect of what?
Happy new year, folks!
Hello, dear reader, and welcome to another issue of AI, Law, and Otter Things! I hope your 2025 has started well. So far, I have little to complain about. After a relaxing break, I feel recharged to move forward with some projects,1 including, of course this newsletter.
Today’s newsletter begins with some developing thoughts on the problem of measuring EU law influence in digital regulation. Then, I move to the usual self-promotion corner, as I highlight recent work outputs and discuss a personal project. After that, I wrap up the newsletter with some reading recommendations and a lovely semiaquatic mustelid.
Measuring influence is hard!
One of my active research interests is trying to make sense of how EU digital regulation affects legislative efforts in Brazil or elsewhere. Nowadays, this discussion is often presented in terms of the Brussels Effect, that is, the idea that EU regulations end up becoming global standards due to a combination of factors. The term “Brussels Effect”, properly understood, refers to a specific mechanism that leads to the spread of EU regulation: under certain conditions, business actors end up complying with EU requirements even outside European jurisdictions, because it makes economic sense to do so. More generally, the term has been used2 (e.g. by Greenleaf) as a metonymy for all the myriad ways in which legislation aroud the world has been influenced by the EU.
Conspicuously absent from many debates on this topic, however, is the side of the jurisdictions that are purportedly influenced by the EU. Engagement with the laws from other jurisdictions is often limited to a superficial examination of textual similarities between the “source” EU legal instrument and the “derivative” law from elsewhere. This is the case both for those who argue for the existence of a Brussels Effect in a certain case (for example, by pointing out similarities between US state privacy laws and the GDPR) and those who argue that a Brussels Effect is not taking place (for example, by highlighting the differences between the Brazilian AI bill and the EU AI Act). I believe this is insufficient in both cases.
First, because non-EU jurisdictions are not merely passive recipients of Brussels’s commandments. Policy scholarship has long pointed out that the reception of “foreign” influences is mediated by local concerns:3 legislators read foreign law templates in light of local political priorities, tweak it to fit the local legal systems, and so on. A good illustration of this is given by Thales Martini Bueno and Renan Gadoni Canaan’s work on the Brazilian fake news bill, where they show that stakeholders looked a lot to the EU but not as a source of ready-made solutions. Instead, reference to EU law provided vaguer inspiration or, in some cases, a more rarefied symbolic value of showing that regulation is not technically impossible.
Second, because textual similarity in itself is not enough to show an immediate influence. Legal systems outside Europe, especially in the so-called Global South, have been heavily influenced by European and American laws. Coming back to the Brazilian example, one can clearly see how that legal system is influenced by bits and pieces of other templates: German and French influences in private law, lots of Italian civil procedure, American influences in competition law, and so on. Those influences, in turn, affect the choices of legal scholars who go on to study abroad and later return to their countries. Because of that, textual similarities can also be the product of convergent evolution—that is, similar legal cultures reacting in similar ways to problems—rather than direct transplantation.
The same structural factors point out to a third shortcoming of superficial text matching. The converse of the previous case also holds: a legal approach can be influenced by another even if the ensuing legal texts look different. For example, even though the Brazilian AI bill is more strongly couched in a fundamental rights language than the EU AI Act, it still draws from the same building blocks: a particular formulation of the regulatory objects in terms of systems and models, the risk regulation toolset, the top-down segmentation of risks, and so on.4 So, long story short, please take a closer look at the “recipient” legal system when making claims of legal transplantation or influence writ large. Doing so might tell a different tale, or at least help you nuance the actual forms of influence.
Self-promotion corner
I did not feel the second semester of 2024 as a high point. I was exhausted from the thesis and felt unable to develop new ideas, and even struggled to wrap up my commitments. Fortunately, I had enough support from my peers and friends to pick myself up and situate myself in my new line of work, so I hope to have new things to share soon. But, at the beginning of this year, I could resume work on some of my active research lines and that felt good.
Recently, I shared two new pre-prints. One, forthcoming in the Handbook on the Global Governance of AI (Furendal & Lundgren, eds, Edward Elgar 2025) deals with how the AI Act might influence global regulation. It also features a short introduction to the legislation itself and key themes in its narrative story. The second pre-print deals instead with the question of technical AI transparency: what is it good for? No, the answer is not “absolutely nothing” this time, but I argue that it is much narrower than one expects. Technical approaches to transparency play important roles, but they cannot build trust in AI by themselves, as their use can only be justified if there is some baseline trust present. Otherwise, there is considerable risk of manipulation by interested actors.
Last week, I once again taught at the Brussels School of Competition’s Executive Master on Law and AI. In my lecture, I offered a deep dive into the AI Act’s regulatory logic, its risk regulation frameworks for AI systems, and its standardization and certification mechanisms. This lecture was paired with a lecture by Tjade Stroband from Microsoft, who covered the framework on general-purpose AI models (including the current debates on the code of practice), the AI Act’s enforcement structure, and its approach to sanctions and fines.
As I send this scheduled newsletter, I am currently participating in the workshop Governing or governed by online platforms? Interdisciplinary perspectives on Constitutionalism, the EU and Corporate Power. Yesterday, I presented a paper that builds on the political implications of my previous work on technology neutrality. Many thanks to Marta Maroni and Mariëlle Wijermars for their invitation, and to the participants for sharp feedback that will be very useful as I shape my book proposal.
A personal project
Every year, I set a personal challenge. I try and learn about something that is both (1) outside my comfort zone and (2) is unlikely to have practical use in my work. By doing so, I don’t want to become an expert (or even a journeyman) in something new. Instead, I want to gain some basic knowledge about something that is interesting enough to distract me from my main interests. This exercise is useful to recharge my batteries after long periods of working on something, especially because the absence of any “git gud” commitment allows me to try my hand at something potentially fun without letting my competitive side take over.
Previous iterations of this project have led me to learn more about things like fashion, music, dance, and basic chemistry. For the last two years, Warhammer has shaped my choice of topics: in 2023 I spent some time learning how to paint my overpriced plastic fellas, and in 2024 I threw myself into kitbashing (that is, the idea of using pieces of other models to customize your units, or even creating an entire model using pieces from others). As for this year, I am still undecided about what I should pursue. Any suggestions would be greatly appreciated!
Reading recommendations
Elaine Fahey, ‘The Evolution of EU–US Cybersecurity Law and Policy: On Drivers of Convergence’ (2024) 46 Journal of European Integration 1073.
Mélanie Gornet and Hélène Herman, ‘A Peek into European Standards Making for AI: Between Geopolitical and Economic Interests’ (2024).
Beatriz Kira and Laura Schertel Mendes, ‘Brazil’s Distinct Path in AI Regulation’ (2024) 1 Journal of AI Law and Regulation 240.
Tobias Liebetrau, ‘Problematising EU Cybersecurity: Exploring How the Single Market Functions as a Security Practice’ (2024) 62 JCMS: Journal of Common Market Studies 705.
Christopher Pollitt and Peter Hupe, ‘Talking About Government: The Role of Magic Concepts’ (2011) 13 Public Management Review 641.
Marta Soprana, ‘Compatibility of Emerging AI Regulation with GATS and TBT: The EU Artificial Intelligence Act’ [2024] Journal of International Economic Law jgae040.
And now, the otter
Thanks for your attention! Hope you found something interesting above, and please consider subscribing if that is the case:
If you’d like to continue conversation in any of these topics—or want to share an event, job opportunity, or paper for me to highlight in future issues—hit “reply” to this email or contact me elsewhere. Hope to see you next time!
Both professional and personal: Now that I am almost done with my Orks, I decided to start a small army of Chaos Space Marines. Looking forward to lots of kitbashing.
Improperly, in my view, as this broadening of the term turns the “Brussels Effect” into a designation of origin for well-known policy diffusion mechanisms rather than a distinctive phenomenon.
This view has, to some extent, found its way to comparative law scholarship. See, e.g., Mathias Siems’s textbook.
This is something that I am developing in my work on regulatory monocultures.