CALL member Marcelo Rodriguez, the Foreign, Comparative and International Law Librarian at the University of Arizona Law School in Tucson, Arizona, recently wrote an article entitled "Ceci N’est Pas Un ChatGPT".
It originally appeared on Slaw.ca on June 8, 2023.
It is republished here with permission of the author.
As I finished teaching my class, Foreign, Comparative and International Legal (FCIL) Research, this past semester, a couple of students asked me about ChatGPT and artificial intelligence (AI).
Given the ubiquitous presence of these topics in everyone’s minds, I should have expected these questions. This is clearly what everyone is talking about and my students are no strangers to these conversations.
As someone who works on legal research with sources in multiple languages and from a wide range of countries, I identify myself as agnostic when it comes to technology. In the end, I decided to share with my students the several times when I had to call another human being or when I had to use a fax to receive the materials I needed as major technological breakthroughs in the FCIL field.
Joke aside, there are clearly some major challenges when you pursue FCIL research regardless of advances in technology and the impact it might have on research. To this day, researchers struggle to grasp the idea that not everything is available online. (The horror!).
Simply put, ChatGPT and the promise of AI run afoul this reality in FCIL research. Regardless of the jurisdiction, timeframe, topic of interest, there is a high chance of probability that some relevant information for your research won’t be available online.
As an informed and skillful researcher, you should be able to overcome these challenges and that’s precisely what we work on in my class. And even if the information is available online, you need to assess and evaluate the information very closely: when was it updated? Is it translated? Who translated it? What are the sources of this information?, etc.
I believe there are a few lessons to be learned from FCIL research which can be applied to the current ChatGPT race.
As I have said before multiple times, having a research strategy is paramount to your research. I recommend my students and all researchers to build a research strategy which is both intentional and flexible. As the well-informed researcher in charge, you should be able to construct a strategy which streamlines your intentions and serves as a map to help you navigate the sources and information you will find along the way.
ChatGPT will not help you do that. I have asked ChatGPT several “how to do research” questions and they all fluctuate between incredibly general to completely wrong.
Spending a few minutes crafting and thinking about the steps, keywords, tools, and sources you will need in your research strategy will help you immensely along your research path.
As mentioned before, another great lesson from FCIL research which can be applied to ChatGPT is that not everything is available online.
Everytime I mention this in front of a classroom or even in my office or zooming in while talking to a researcher, I get a glaring look into the void.
Yes, my friends, it’s 2023 and still to this day, there are big chunks of legal information simply not available online. As someone who specializes in legal systems from around the world and internationally, I encounter this challenge regularly.
However, I also know from experience that this situation also arises in several developed countries, where legal information from lower courts is simply not available in major legal research platforms.
As it has been reported several times, if ChatGPT can’t find the information online, sometimes it makes things up: fictional cases, legislation, articles, etc.
In the FCIL world, if we can’t find the information we need online, we rely on each other. FCIL librarians have developed networks, groups and list servs where information is shared and other law librarians from all over the world help each out when needed.
By mentioning some FCIL lessons, my intention is for all of us to conceive ChatGPT as a tool, indeed a powerful one depending on what you need.
However, it’s still just a tool among many others empowering researchers to do their work efficiently.
As an educator, I’m a big believer in sharing information, building transparent pipelines and allowing people to think critically and rationally about the steps they are taking in an informed way.