The uncommon case of common specifications
Thinking aloud about niche interests
Hello, dear reader, and welcome to another issue of AI, Law, and Otter Things! As you might have inferred from the sort of regular update scale these days, I’m currently procrastinating some non-urgent work trying to make sense of some strands of literature I’ve been recently exposed to. However, being the reflection-averse mustelid enthusiast that I am, I usually find it easier to think when doing stuff.1 So, I will try and make the best of the Boimler Effect to think aloud here with you, and hopefully not embarrass myself too much in the process.
Today’s newsletter follows the usual structure. First, I jot down some words on a niche bit of EU digital law: common specifications, the awkward sibling of harmonized technical standards. Then, I share a few things you might find interesting, followed by some open calls and other opportunities. Finally, as usual, an adorable otter.2
But, before that, a question: are you an EU law scholar who did some of your legal and/or graduate education outside the Union?3 If so, please don’t hesitate to reach out. I am trying to put together some (in-person or online) space for discussing the experience of learning EU law as a foreign language.
What are common specifications, after all?
In the debates surrounding the AI Act and its implementation, a lot of attention has been dedicated to the role of harmonized technical standards (hENs). I have written about that with Nicolas Petit, and we were far from the only ones: Olia Kanevskaia and her coauthors, Marta Cantero Gamito and Chris Marsden, Mélanie Gornet and Winston Maxwell and Hans Micklitz are just some of the scholars who cast light into how the reliance on standardization to provide key aspects of the Act’s provisions raises legal and political questions rather than being just a boring technical affair. Today, I want to raise some questions about the role of a related instrument: common specifications (CSs).
My first engagement with this particular instrument comes from the AI Act, which in Article 41 empowers the Commission to adopt, under certain circumstances, implementing acts supplying technical specifications. Given the exceptional character of those instruments, I did not spare them much thought back them. And that seems to be the case for most of the literature, which usually limits itself to mentioning their existence before focusing on the manifold implications of hESs. This is understandable, given that standardization for the AI Act is already underway and few other EU legal instruments feature CSs. But, after seeing these specifications appear again in some other instruments, such as the machinery regulation, a regulation on the internal market during emergencies, and the Cyber Resilience Act, I wonder if there is more to this instrument than it meets the eye.
In a substantive sense, CSs are similar to hENs, as they both offer technical guidance on how a product can meet legal requirements of quality. At first glance, this might suggest a throwback to the “old” approach of EU product safety legislation, in which the law goes beyond defining “essential elements” to supplying more detailed instructions. However, these specifications bear another important similarity to hENs: in both cases, compliance is voluntary, as providers remain able—at least in theory—to comply with the law by meeting the essential elements in another way. So, CSs provide a departure from the market-driven construction of hENs, but not one as drastic as to restore a command-and-control structure.
The main question that brings my mind back to these instruments is one of purpose. Why has the Commission seen fit to introduce this instrument into recent product safety legislation? After all, the New Approach to product safety is often defended in terms of allowing regulation to tap into expertise that would not be readily available to the EU institutions. Indeed, there has been some scepticism about whether those instruments will be used at all, with some experts suggesting that common specifications might end up being outsourced to consultants and/or incorporating bits and pieces of incomplete technical standards. This seems to be a plausible scenario, even if one that is somewhat difficult to reconcile with the Commission’s decision to include CSs into its regulatory toolbox both before the AI Act4 and in the posterior instruments mentioned above.
Accordingly, I am curious about what role those instruments are expected to play in the regulatory architecture of the Single Market. The few works that discuss CSs more at length, to my knowledge, focus on its deterrent role of shaping the behaviour of European Standardization Organizations. The regulation on emergencies mentioned above suggests also that non-economic factors, such as responses to crises or even broader national security concerns, may also be invoked to override (or at least tweak) the market-based approach to standardization. And, on a rosier note, my work with Petit has also explored (though not exactly endorsed) the possibility that CSs emitted by the Commission might avoid some of the legitimacy gaps of hENs when dealing with the “fuzzier” aspects of public values protected by the AI Act through standardization. It remains to be seen whether and how those policies might push the Commission towards taking a more active role in technical rulemaking, at least from a formal perspective.
If it does, then we might find ourselves needing to spend more time with the formal and substantive issues that arise out of this instrument. Many of them might be cast in terms of established procedures, in particular because the specifications need to be adopted through an implementing act (hence, the need to think about comitology). Yet, we might find out that — in addition to the peculiarities of specific regulatory domains such as AI — there are also some practical aspects to be ironed out about the instrument itself. As of the moment, I am not really in a position to offer clear answers to the speculatory questions posed above, but I am working on them in background mode, especially since the security angle is relevant for my work in the Chair in Cyber Policy. Please don’t hesitate to reach out if you happen to be working on CSs or on related questions!
Recommendations
Deviating from my areas of competence, Projeto Ariranhas (a group focused on the preservation of the Giant Otter) has just released a report on the conservation status of the Pteronura brasiliensis in various South American countries. The report is available in Portuguese and English and, among many news and interesting facts, it discusses some sightings of this species in Argentina, where it was thought to be extinct for a while now.
As for works on law, AI, and related areas, here are some suggestions:
Mark Dawson, ‘The Changing Substance of European Law’ (2024) 20 European Constitutional Law Review 451.
Emilia Korkea-aho and Päivi Leino, ‘INTERVIEWING LAWYERS: A CRITICAL SELF-REFLECTION ON EXPERT INTERVIEWS AS A METHOD OF EU LEGAL RESEARCH’ [2019] European Journal of Legal Studies 17.
Miikka Hiltunen, ‘Exploring the Legal Making of the European Digital Economy’ (2024) 3 European Law Open 633.
Bryan McMahon, ‘Bubble Trouble’ (The American Prospect, 25 March 2025).
OECD, ‘Recommendation on Public Policy Evaluation. Implementation Toolkit’ (OECD 2025).
Annalisa Pelizza and others, ‘Manifesto on the Datafication of Mobility Across Borders’ (2024) 10 Engaging Science, Technology, and Society 207.
Ari Ezra Waldman, ‘Challenging Technology Expertise’ (Social Science Research Network, 1 January 2025).
Self-promotion corner
Next Friday (4 April), I will present some earlier outputs of my work on regulatory monocultures at BILETA 2025 in London. Don’t hesitate to drop by!
A few days later, I will talk about the legal sources of technical opacity at XXII Jean Monnet Seminar “Modern Technologies and EU Law” in Dubrovnik.
Also, I am proud to say that I am now a member of the Center for AI and Digital Policy (CAIDP)’s Global Academic Network. CAIDP does an excellent job of raising awareness, training people, and actively engaging with matters of AI and democratic values, and it is a honor to join some amazing people in the network and contribute with this mission.
Opportunities
On 3 April, CAIDP will host a roundtable on AI and Democratic values. Check it out!
The COMPASS workshop at ICWSM 2025 (AAAI Conference on Web and Social Media) invites multidisciplinary contribution on content moderation until 7 May. The event will take place in Copenhagen on 23 June.
Also in the UK, LSE Law is hiring two postdoctoral Fellows. Apply by 27 April.
The second annual conference of the EULEN network will take place on 11 and 12 December in Utrecht. They invite papers until 1 May.
King’s College London seems to be hiring a large cohort of research fellows dedicated to AI-related topics across disciplines, with a 5-year contract that potentially transitions to open-ended academic employment. Apply by 19 May.
It’s just one thing after the otter
Thanks for your attention! Hope you found something interesting above, and please consider subscribing if you haven’t done so already:
Do not hesitate to hit “reply” to this email or contact me elsewhere to discuss some topic I raise in the newsletter. Likewise, let me know if there is a job opening, event, or publication that might be of interest to me or to the readers of this newsletter. Hope to see you next time!
Incidentally, this is the reason I never got a driver’s license and have no intention of doing so in the future.
Attentive readers have noted the absence of Warhammer miniatures this time. This is because I had no time to paint my Alpha Legionaries. However, I set up a permanent space for painting minis at home, which means I can at least free up some space in the dining table. And, with the sun finally deigning to appear around these parts, I can go back to priming minis with sprays, saving me some time.
While exchanges are certainly a valuable tool for broadening one’s mind and education, I am thinking here of longer passages, such as an entire degree or at least a co-tutelle process.
To my knowledge, the Medical Devices Regulation was the first instrument to bring the possibility of common specifications. It has also led to the adoption of some instruments using those powers, though one also see mentions to “common specifications” in the context of railways (which are covered by a different regulatory approach) and CBAM.