Lawyer-programmers and online mustelids
Welcome to the second issue of AI, Law, and Otter Things! As promised, today’s newsletter will spend some time on my favourite mustelids and how they reached that position. We’ll also talk a bit about programming and the law, and I’ll close this newsletter with some recommendations of things I enjoyed.
(Feel free to put some upbeat music on before moving on to the rest of the newsletter.)
Should tech lawyers learn to code?
Over the last few weeks, a lot of my time went into organising AI and Law Summer School, hosted by the EUI and the University of Pittsburgh. My role in this event was primarily administrative, but I had the opportunity to follow lectures by various researchers working on applying artificial intelligence techniques to the legal domain. It was an exciting event, covering both machine learning and knowledge representation methods, and the online format allowed us to have students from various backgrounds and regions, enriching the debates during the event.
An interested reader might have noticed that the Summer School programme does not include any classes in computer programming. While some of the lectures included hands-on experiences or guided tours of legal AI software, the school’s main goal was to present a high-level view of the current state of the art and its limitations. Such a perspective allows students to understand what is going on, distinguish feasible claims from misleading sales pitches, and engage with technical stakeholders in developing and regulating such applications.
I firmly believe that this is the way to go when it comes to technical knowledge for lawyers. Fortunately, the “all lawyers need to program” discourse, just like its general counterpart, seems to have been replaced by more nuanced variants, which argue that programming might be helpful for some lawyers. These lawyers would benefit from programming either by developing software (on their own or within a team) or obtaining insight into how computer systems are made. This second reason, in my opinion, is somewhat misleading.
Of course, I am not saying that my background as a computer scientist does not inform how I look at AI systems and their regulation. But claims that (some) lawyers must learn how to program often overlook two important facts. Since programming is in a field with its disciplines and specialised knowledge, acquiring more than a passing familiarity with programming requires substantial time and attention. Even if a lawyer wants to invest those scarce resources, they must be careful while navigating between the Scylla of excessive trust in their own knowledge[1] and the Charybdis of subordinating legal concerns to technical considerations.[2] A programming-informed understanding of the law requires a lot of effort to avoid cognitive traps.
To make things worse, these efforts might have a limited payoff. Even though there are differences between how an object is seen by its creator and external observers,[3] programming is only a fraction of the broader field of computer science. Looking at a software system from the eyes of a programmer presents an incomplete picture of that system, and the resulting gaps may distort our understanding, for example, by omitting the role of the technological infrastructure.[4] At this point, a bird’s-eye view of what is going on in a computer system might be more informative than the insider perspective of a programmer.
So, if you are interested in automating work or in the intellectual pleasures of programming, then by all means learn how to program.[5] If you are not sure programming might be for you, give it a try: many people find it an entertaining activity.[6] But, if your interest lies in understanding what is going on, it might be a better idea to learn about the general features of computer systems and collaborate with computer scientists.
[1] My first thought here was the Dunning-Kruger effect, but it turns out that the meme version of this effect does not really hold, as often is the case.
[2] Mireille Hildebrandt has an interesting discussion of interdisciplinary work in tech law.
[3] For a pre-computational example, see Giambattista Vico’s La scienza nuova.
[4] See, e.g., Ben Allen, ‘Source Code Isn’t’ in Thomas S Mullaney and others (eds), Your Computer Is on Fire (The MIT Press 2021).
[5] Thibault Schrepel, who is more open than I am to the entire lawyer-programmer enterprise, has compiled a nice list of learning resources.
[6] Especially if you like puzzles.
You had promised me otters...
In the previous newsletter, I had mentioned that I would explain why otters will be such a frequent theme in this newsletter. Well, the most obvious reason is that they are adorable:
They are semi-aquatic animals that have adorable looks and are relatively intelligent. What is not to love? I could — and probably will, at some point — spend an entire newsletter speaking about these lovely mustelids.
But this does not really explain why they have become part of my online brand, does it? The answer to this question is actually related to another animal:
Given that I started using the internet in the early 1990s, my online education comes from a time before real-name policies. Back then, it was acceptable — even expected! — that one would make themselves known online through a relatively stable online handle, which might or might not be related to one’s name in “meatspace”. For better or worse, pseudonymity has been very important to my growth as a human being, and I have tried to maintain it in my online presence as much as possible. And this is where otters come into play.
Some years ago, I was looking for a new online persona. Not only my previous profile had been leaked at work — creating no real problems, but nevertheless making me too self-aware —, but I had grown out of it. So, it seemed the ideal time to create a new handle. At the time, I was reading a Tumblr name Discourse on the Otter, which combined theory puns and cute otter photos, and so I decided to try to use an otter avatar. Hilarity ensued, and I progressively built a new online persona around the mustelid theme.
My online presence as a pseudonymous otter has been surprisingly fun. It allowed me to make lots of friends (including my wife!) and introduced me to relevant people in my area who deigned to discuss stuff with a swimming mustelid. Even though I eventually relaxed the separation between my personal and professional identities, maintaining different profiles is still helpful for audience segmentation. And, as long as I keep separate profiles, I might as well keep the cute otters as a distinguishing element of my online presence. So, expect to see lots of adorable semi-aquatic mammals in every edition of this newsletter.
Recommendations
Given that I wrote more than I expected in the previous sections, I will try to keep my recommendations brief:
Language learning app with an interesting approach to community-based correction of written and spoken production.
Gwyneth Cheng, ‘Wild Otters: Threatened and Traded’ Kontinentalist (2021)
Engaging journalism on the illegal otter trade in Asia.
TIE Fighter Total Conversion for X-Wing Alliance
A recently released game mod recreates the 1994 space simulator in the XWA engine, as improved by other fan-made mods. It includes all campaigns from the original game and a partial release of reimagined versions of these campaigns that make use of the available technological resources.
By the time you’re reading this newsletter, I will probably be in Rome for my first medium-distance trip after moving to Italy! Hopefully, this will give me some exciting topics for the following newsletter.
See you, space cowboy…