AI has arrived, and it’s easier than ever to produce summaries of cases, research articles, or anything else that’s hard to read. So, why bother reading at all?
I’m not against using summaries or other tools to help us read the full version of something. Even before AI, lawyers and law academics used summaries of cases published in newsletters and on websites and blogs to stay on top of legal developments, understand what their implications might be, and to decide which cases to read in full. AI has become another useful way of summarising and analysing legal material, particularly as it becomes more sophisticated and able to produce more accurate summaries.
But why read a text in full? One obvious point is that both humans and AI make mistakes. AI can also produce hallucinations (such as citations to fake cases), although these are less common in human-made work (unless you’re reading something written by Hunter S. Thompson).
Beyond this, all texts are open to interpretation, and the interpretations produced by AI or by another human may not be the same as the interpretation we produce ourselves. We need to make sure that we understand a case or an article or a book, and that we can produce our own interpretation of it. It’s particularly difficult when you’re starting out, but I encourage students (and anyone at any stage of their career) to reflect critically on any writing produced by someone/something else and challenge whether it is accurate, logical, and defendable.
Sometimes, the process of reading means writing your own summary or reflection on what you’re reading. For many people, this is a critical part of understanding what you’re reading and producing your own thoughts on it. For this reason, there is pushback on using AI in the writing process, because it interferes with the thinking process (and for other reasons). Too much reliance on human or AI-produced summaries may interfere with the reading-thinking-writing cycle.
In the domain of law (at least in common law legal systems), there is a foundational distinction between the common law, i.e., the legal principles developed by judges when deciding cases, and statute law, i.e., laws passed by parliament. Judgments (usually) follow a particular structure that students and practitioners need to understand: judges set out the facts of the case, identify the issues the case is concerned with, describe the relevant legal principles or statutory provisions, and apply these legal principles/provisions to the facts, in order to come to a conclusion about how the case should be decided. The development of new legal principles occurs when there is uncertainty or a gap in the law that needs to be applied in a case, which the judge resolves through developing a new common law principle, or deciding how a word in a statute should be interpreted. In common law areas of the law (such as tort law), we don’t simply read a case in isolation: we situate a case within a line of cases to understand how a legal principle or test has developed over time.
It’s crucial that law students understand the basic structure used in (most) judgments, and that they can identify the different parts of a case. This helps us understand the critical issue(s) the case is deciding, and it speeds up reading cases, as we understand what each part of the judgment is doing, and what sections we need to read more closely, as compared to those we can skim through. Students also use this issues-rules-application-conclusion format in answering legal problem questions (or engaging in ‘issue spotting’), so understanding this form of reasoning is fundamental to answering different types of law assignments. Students (and practitioners and law academics) need to read a sufficient number of cases, from scratch and without assistance, to be able to master the basic skills involved in identifying the different components of a case, the issue(s) the case is deciding, and the new legal principle(s) decided by the judge (if any). For now, we need to know how to read cases ‘manually’ first, before using tools to assist us. This form of case analysis is a hallmark of the legal discipline, as well as a fundamental skill of a competent practicing lawyer.
I’ll leave statutory interpretation to one side, but equally, it involves specialist skills and reading techniques (those without a law degree may not realise this), which need to be developed ‘from scratch’ before using electronic, algorithmic, or AI tools in statutory analysis. The comments I’ve made above about cases also apply to legislation.
In the domain of research, a similar situation exists, in that abstracts provide us with a summary of a research paper, but we need to read a paper in full for a comprehensive and accurate understanding of what it says. But there are other reasons for reading research articles (in full), other than to understand a new perspective or new data on a topic. Researchers often do their best work as part of a team, when they can pool their knowledge, bounce ideas off each other, or challenge each other’s approach. Reading other people’s research can also give you insights into what methodology to use in your own study or how to push your research further, or what considerations or concepts might be missing.
Research publications are not just a form of knowledge dissemination. They are an important part of the dialogue between researchers, and the creation of a distinct research field. Perhaps more importantly, they are a way for researchers to be in community. I work in the field of public health law, which is relatively small in Australia, but is a much larger, well-established field in the US. I read articles published by US public health lawyers to understand how the field is developing there and what US academics are concerned about. An AI summary of Professor Scott Burris’ article on ‘What is a public health lawyer today’ would probably not reflect his concerns about the field and his pursuit of social justice – using laws that end poverty and discrimination, and provide basic goods and services like parks, transport, and education, which are the most fundamental way to improve our health. An AI summary would also probably omit some of the fruitier footnotes – or maybe I’d assume they were hallucinations?
We are overwhelmed with a massive amount of material to read. I could read nothing but tort law cases for a year and still not be up to date. There is nothing wrong with using summaries and other tools to help us quickly understand new developments or new knowledge in an area, and to decide on what we do need to read in full. However, reading is critical for developing foundational legal and research skills, and for engaging in the dialogue and intellectual exchange that holds academic disciplines together.
Leave a Reply