Hello, dear reader! Today’s update is a short one. Between my thesis, some pending tasks, and Ace Combat 7, I did not have much time to write the essay I had in mind for today’s issue. Or to finish it by Wednesday as originally intended. Nonetheless, I felt that a quick update would be a good way to end this rainy Florentine Friday. So, the following lines will mostly feature recommendations. And the usual otter, of course.
Before moving on to that, however, I am legally obliged to discuss the latest AI Act drama. In a previous issue, I complained about how the trilogues seemed poised to introduce the new category of “very capable foundation models”, which added some problems to those already connected with the legislative use of the concept of foundation model itself. According to the latest news from Euractiv, there is less consensus on this matter than I feared.
Some countries—mainly Germany, France, and Italy—seem to be pushing for the AI Act to not include any rules on foundation models at all, as such rules would potentially harm their putative national champions. The Parliament negotiators seem unhappy with that development, as the special regime for said models was one of the key features of their proposed amendments. As a result, there is a chance that the trilogues will not reach an agreement before the political meeting on 6 December 2023.
If agreement is not found in the next few days, reports Euractiv’s Luca Bertuzzi, the whole AI Act might be at risk. This is because next year is an election year for the European Parliament, one of the EU’s co-legislator institutions. If there is no political agreement on the contents of the law by the end of this year (or at most early 2024) there might not be enough time to complete the formal legislative process before the end of the current mandate of European parliamentarians.1
It might be a bit premature to suggest that the AI Act will have the same fate as the ePrivacy Regulation,2 languishing in a limbo without ever being formally rejected. Still, as Bertuzzi and other EU insiders point out, there seems to be precious little time for finding an agreement in Brussels. Especially after the US executive order on AI, which not only steals a bit of the EU thunder but weakens the AI Act’s claim to influence as a regulatory first mover. Therefore, the end of the year might be a bit more thrilling than expected for those following AI regulation in the EU and beyond.
Recommendations
It has been a while since I’ve last recommended sci-fi stuff around here. Today we had the first episode of For All Mankind season 4, but it is still a bit too early to recommend it (especially after the mess that was season 3). Instead, I will suggest two old sci-fi short stories that you might read over the weekend. Arthur C Clarke’s Superiority is a fun take on technological solutionism and the problems it creates. As for the other one, David Gordon’s Despoilers of the Golden Empire I will break my pro-spoiler rule and refrain from telling you anything about it, but it is an interesting exercise in metafiction.
As for more academic indications, you might be interested in the following:
Marco Almada, ‘Governing the Black Box of Artificial Intelligence’ (SSRN Scholarly Paper, 7 November 2023).
William Nichol Eskridge and others, ‘Textualism’s Defining Moment’ (2023) 123 Colum L Rev 1611.
Nora von Ingersleben-Seip, ‘Competition and Cooperation in Artificial Intelligence Standard Setting: Explaining Emergent Patterns’ (2023) 40 Review of Policy Research 781.
Alexandre Lodie, ‘Are Personal Data Always Personal? Case T-557/20 SRB v. EDPS or When the Qualification of Data Depends on Who Holds Them’ (European Law Blog, 11 July 2023)
Matthijs M Maas, ‘AI is Like… A Literature Review of AI Metaphors and Why They Matter for Policy’ (Legal Priorities Project, 2023).
Virginia Passalacqua and Francesco Costamagna, ‘The Law and Facts of the Preliminary Reference Procedure: A Critical Assessment of the EU Court of Justice’s Source of Knowledge’ (2023) 2 European Law Open 322.
Manon Revel, ‘How to Open Representative Democracy to the Future?’ [2023] European Journal of Risk Regulation FirstView.
Sarah Tas, ‘The Dangerous Increasing Support of Europol in National Criminal Investigations: An Additional Layer of Complexity’ [2023] New Journal of European Criminal Law 20322844231214482.
You should also check the open template syllabus Kevin Frazier is preparing on Regulating AI: Legal and Policy Perspectives.
Finally, academic readers might be interested in two conferences in the US:
Michigan Law’s Junior Scholars Conference, with the theme Reimagining the Public-Private Divide. They are accepting abstracts until 5 January 2024, with final papers due by 31 March 2024, and the event will take place on 12–13 April 2024.
Also at the University of Michigan, the Legal Philosophy Workshop invites abstracts until 15 January 2024 for an event in 13–14 June 2024.
That’s it for today! Hope you have a lovely weekend, and please don’t hesitate to get in touch with any comments or suggestions.
Followed shortly after by the end of the term of the current European Commission, which might lead to changes in policy.
Is that the worst-case scenario for the Act, however?