I was recently speaking with someone about the Internet’s “Eternal September”, which is the concept that starting in 1993/1994 so many new users started using the internet that it could never settle into the online equivalent of November on campus, when people have figured things out where their classes are and where to get the sandwiches they like for lunch. From the perspective of 30+ years later, with no indication that the phenomenon has slowed, wearing a t-shirt saying that the internet was full in 1993 seems to lack, if not imagination, then at least prescience. However, the cultural effects of a constant influx of new participants over this entire period has had limiting effects on the quality and development of certain systems (Ivan Vendrov explained this well in “The Tyranny of the Marginal User“, writing that the peak for online dating was OKCupid in 2016). So, we have systems that have to be engineered for an extensive range of skill levels, and quality and age of equipment, which limits our ability to build applications and services that would benefit those who put in the time and effort to become proficient.
Beyond the internet, this seems to have clear parallels with the adoption of technology and quantitative methods in the legal sector. For example, legal AI has been developing since the 1970s, and the quantitative study of legal topics has been happening for decades. As these systems and bodies of knowledge mature, and the culture shifts to adopt them more widely, new users have the potential to enrich the community. But developing functional ways to help them understand what’s possible and integrate them into systems has the potential to make this process go more smoothly — I have watched people’s eyes fill with wonder when taught to do an online search or use ctrl-f.
Everyone brings their hobby horses with them through life, and as I turn my focus to legal scholarship, I see that there are significant gaps in the ways these approaches have been used. There are many areas of research that have made extensive use of these techniques, such as criminology, economic analyses of law, tax, and others. However, there are also many gaps in how it has been explored in other topics, and the increased calls for the use of data in more areas of public administration and research has been an important development in recent decades.
There are many reasons why there haven’t been as many projects that integrate data science into legal contexts as some may prefer. Often there are roots for this in technological limitations: until recently computers weren’t powerful enough to process complex language. These reasons are also related to cultural preferences, as many people who choose to pursue legal careers are not as comfortable with quantitative methodologies as they are with qualitative approaches. Developing sufficient expertise to envision what research can be done and learning to do this research have significant learning curves. That said, there are so many opportunities for meaningful research in most, if not all, aspects of legal scholarship.
These limitations are shifting as computers continue to become more powerful and the legal culture shifts to endorse them. The technical limitations are also disappearing, as tools develop sufficiently to be used by people with lower technical skill levels. However, we are still in the early stages of this change. This means that for legal scholars and practitioners there are many opportunities to research questions surrounding issues like how people experience the legal system and what impacts different policy changes may make.
Many researchers have spent some part of recent years looking at how systems such as large language models can be used in law and how they can be made more reliable. There are many potential uses for this work, and it is promising for innovating legal clinics, practice of law, governance, and understanding how the law works in society. The availability of large language models means that we are going to be living in a more data rich environment as unstructured and semi-structured data is easier to transform into more computationally tractable formats.
One of the topics that’s most commonly brought up in relation to data driven research in Canadian law is the lack of accessible case law, but there is also a lack of data on many other aspects of our legal system. This means that not only do we miss having case law available for analysis, which gets so much attention. We also don’t have access to many other sorts of data that could be used to understand justice in our society. There are countless opportunities to use these methodologies, and it seems like a missed opportunity for lawyers and others in the sector to not be included in key decision making due to issues like a lack of understanding of what’s being discussed or misaligned priorities within project teams.
Though it has been gaining momentum for at least a decade, we are in the early stages of this change. And there are still significant opportunities for finding quick wins using existing models. There are also meaningful prospects for data experts to do novel work with real world applications on challenging datasets. The combination of these conditions means that this is a perfect time for data scientists and legal scholars to explore how they can best work together. As more people approach the study of quantitative approaches to law, it can be hoped that in ten years we won’t read about the greatest time for data analysis in the sector being 2025. Having spent so much time in school, scholars know what to do in September to ensure that they and others move toward. I look forward to the progress toward a more general understanding of the potential for these approaches and how this work progresses.
I would like to thank Burkhard Schafer for discussing the content of this column with me.
The post The Opportunities Hidden in Law’s Eternal September appeared first on Slaw.