While there is growing interest in incorporating artificial intelligence (AI) into the legal research and writing curriculum, little agreement exists among instructors or law schools on how to do so. In recent years, the need to address this gap has been bolstered by the adoption of the rule for duty of technological competence in the United States and Canada. Yet there is no real standard for introducing new and emerging legal technologies into the curriculum or course offerings for law schools.
One recent survey of American law schools found the following: six schools offer courses on artificial intelligence, approximately 15 offer a legal technology certificate, over 40 have clinics or legal technology labs, and over 70 offer law and legal technology practice management courses. Even the high end of this range represents fewer than half of all ABA-accredited law schools and reflects some of the vast differences in how technology is (or is not) included in a law school education more generally.
Various authors and practitioners have specifically noted the need to incorporate the competent use of AI into the law school curriculum. But the sheer number of technologies available in the legal industry makes this an enormous task. How can instructors possibly teach students about every new or emerging legal technology?
In this article, I discuss how information literacy provides a useful framework for integrating AI-driven tools into the law school curriculum.
I. Artificial Intelligence and Legal Education: The Basics
When we talk about incorporating AI-driven tools into the law school curriculum, we are talking about a specific subset of emerging legal technologies. “Artificial intelligence” is commonly defined as any technology that allows a machine to replicate tasks typically associated with higher level human cognitive processes, such as learning from experience and understanding language. Emerging legal research and writing tools have leveraged AI technology to facilitate research and writing tasks in many ways. Often, AI appears as an enhancement to an existing or improved platform, such as enhanced search algorithms integrated into major legal databases. Other examples represent a much more dramatic attempt to automate an entire skill or behavior, like an AI-driven tool that generates an entire research memo.
No matter the extent of dependence on AI, however, these emerging technologies all share one commonality of concern to legal research and writing professors: they place more of the onus of research and writing on the design and capabilities of the machine, while the researcher is removed to some degree from the process. As a result, these emerging technologies all limit the user’s ability to control, understand, and adjust the direction and results of the research process. If these tools are not taught alongside basic research skills in law school, students encountering them in the workforce will lack the basic understandings required to integrate them into their practice.
The “practice-ready graduate” is a concept that arises frequently in discussions about how to teach technology in law school. This means students “should have received an introduction to broad doctrinal precepts and the basic tools needed for access, responsible use, development, and comprehension of the law.” The focal point on “tools” in this definition is significant. It refers to a belief that a foundational understanding of the law is not enough for legal practitioners. In the context of the duty of technological competence, these expectations extend to an assumption that students should be well-equipped to work with new and emerging technologies, including AI-driven tools.
The challenge with teaching emerging technologies is, of course, that the landscape can change dramatically in an exceptionally short time span. A legal tech tool that is available one year may very well have shifted focus, changed drastically, or completely failed within the span of that same year. Other technologies that were not readily available one year may see a quick uptake or be made accessible to students and graduates in similarly short intervals. Furthermore, there are so many practice-specific technologies by now that it would be a fundamentally impossible task to equip every student with knowledge of the technologies needed to practice in the specific areas of law most relevant to them.
Consequently, teaching legal technologies with a focus on specific tools is extremely difficult. The landscape changes far too quickly to give educators the chance to adequately integrate them into the curriculum. As noted by Margolis and Murray, this rapid pace of change is a characteristic of the legal landscape that is crucial to teaching legal research and writing more broadly speaking:
Legal educators […] must move beyond an understanding of research and writing as fixed skills with clearly defined parameters. While the fundamentals of research and writing are still important, it is time to start broadening our understanding of skills to include the ability to self-learn, to ask the right questions, and to evaluate and incorporate new methods into existing skillsets.
This is where the concept of information literacy is helpful. The following section will describe how instructors can use this lens to introduce AI competencies into existing course offerings.
II. Artificial Intelligence Through an Information Literacy Lens
Information literacy is a concept that centers instructional efforts on improving how researchers interact with information itself, instead of with specific tools or resource types. It is “the set of integrated abilities encompassing the reflective discovery of information, the understanding of how information is produced and valued, and the use of information in creating new knowledge and participating ethically in communities of learning.” As such, it is practice- and course-agnostic. This makes it a helpful framework for teaching about technology because it centers the information at the heart of those technologies instead of the tools themselves; it is not tied to a specific technology or modality of research and writing. In the legal industry, where old and new modalities coexist in seeming perpetuity, this is an especially helpful frame. Instructors need to teach students to navigate print-based and AI-driven research at the same time.
In its clearest form, there are five fundamental competencies described by the American Library Association (ALA) that students should be able to demonstrate in accordance with an information literacy approach. These include being able to:
know when and to what extent information is needed;
access information effectively and efficiently;
evaluate information critically;
use information for a specific purpose; and
understand the ethical, legal, and social issues surrounding information.
While this framework has been developed more broadly for the work of library and information specialists, it has also been adapted to meet the needs of specific disciplines, including law. Table 1 below shows the ALA standards alongside the American Association of Law Libraries’ (AALL) Principles and Standards for Legal Research, to illustrate how legal information specialists already incorporate these principles into educational design.
There is much more detail in the AALL document that breaks down these principles into specific competencies, including several that speak directly to the emergence of new technologies. But these five identified principles are “broad statements of foundational, enduring values related to skilled legal research.” And their very broadness is what gives them continued relevance when applied to AI-driven technologies.
Using this framework to interrogate AI-driven legal research tools is logical because the components of these new tools all fundamentally rely on information. As experts note, an AI system is composed of three relatively distinct aspects:
the input (or data) that the system draws from and operates on;
the algorithm(s), or the reasoning, that is applied to that dataset; and
the output that the system produces.
Let’s take as a simple example an AI-driven case law search. The input for this tool is the database(s) of case law on which the tool runs.
The algorithm(s) is how the tool has been programmed to interpret that data. This may include explicit programming (for example, to prefer decisions that are more recent or to contain more repetitions of the search terms). Many AI algorithms, however, are a black box to even their creators. For instance, with the subfield of AI called machine learning, the computer discerns patterns from data that are not perceptible to humans and “learns” to apply those patterns moving forward without ever providing insight into why it is reaching those conclusions.
The output for this tool is likely a list of results that provides the researcher with potentially relevant judicial decisions.
While there are certainly aspects of an AI-driven system that are unique from technologies of the past, the basis of all three components here is still information. This makes it relatively simple to parse into an information literacy framework. In Table 2 below, I use the information literacy standards as a basis for adapting the above AALL principles to specific legal skills related to AI. For educators, this can help establish learning outcomes when incorporating AI into a class or course.
This exercise, while by no means exhaustive, is a useful consideration for anyone looking to incorporate AI into the legal research and writing classroom because it provides competencies that can be used as the basis for backward design.
For example, if I wanted to target the “evaluate” standard in an educational session, I could use these as the basis for my learning outcomes:
To meet those outcomes, I may design a session with the end goal of having students learn and apply a framework for evaluating a brief analysis tool. I may structure this lesson by first providing the evaluation framework based on the criteria listed above: relevance, authority, credibility, currency, and bias. Then, I could have students test out a brief analysis tool by uploading sample documents, such as two legal briefs on the same topic. What happens when students compare the results when the tool has interpreted two documents—are there similarities? Differences? Why are the results not identical?
This approach is very high level and focuses on the information at the heart of the tools. There is little detail on the actual law itself. By using information literacy as a framework for evaluation, we target skills and practices that will be relevant to students no matter in what area of law they end up practicing. The focus is on the tool as point of access and use of information in the broader sense and not on the specific strengths or weaknesses of a particular technology.
These are all skills that are not tied to a specific area of law, build on each other, and cannot be learned in a single session. As described by the Association of College and Research Libraries, information literacy should be viewed as “extending the arc of learning throughout students’ academic careers and as converging with other academic and social learning goals.” In an ideal world, therefore, information literacy is fostered across the curriculum of a given program and not simply confined to sessions offered by the law library or to legal research and writing courses.
III. Looking Across the Curriculum
How do we integrate this type of training across the curriculum in a way that exposes students to the tools and skills necessary to succeed?
The law library has some specific strengths and weaknesses in initiating the integration of these skills across the JD program. First, it is helpful that librarians already have the skillset required to teach information literacy sessions across the curriculum. This makes them ideally positioned to take on this type of instruction, even though some may find AI to be an intimidating topic. Librarians and library staff also have a higher vantage point than individual instructors, who may find themselves limited by the specific subject matter that they teach.
There are, however, also key limitations. It can be difficult for librarians to make inroads to instructors and students, who may not be aware of this skillset or the fact that librarians can teach these types of skills. For example, several studies have found that both faculty and students are not always aware that information literacy instruction is part of an academic librarian’s role. It may also be difficult to provide students with access to certain tools and technologies, depending on the law library’s budget and geographical location. Not all vendors are willing to provide students with free access to an AI-driven tool or are responsive to inquiries from academic faculty and librarians. Other tools are created for a specific market (most often the U.S. legal market) and are therefore difficult to use as a demonstration tool when teaching in another jurisdiction. Additional barriers may include lack of reliable internet access for some students, including those in remote locations.
An information-literacy driven approach to teaching AI from the law library therefore has additional layers beyond the approach outlined above. First, tools must be identified that can be provided as a method of bringing AI to life in the classroom. While sessions themselves will be designed with learning outcomes based on information literacy, instructors must provide students access to tools and technologies first to begin building programming. Leveraging existing access, such as technologies that are integrated into major legal research platforms, is a useful starting point. This approach both saves time that would otherwise be spent on identifying and obtaining access to new tools, and also increases the possibility that student learning will be reinforced later on through use, as they are much more likely to encounter common tools such as Westlaw Edge and Lexis Advance Quicklaw throughout their professional and academic careers. Vendor outreach can also be fruitful, however, as some companies are willing to provide access to students for use in an educational environment. Benefits of reaching out to smaller vendors include increasing student exposure to more types of AI-driven tools beyond those offered by the biggest vendors and supporting legal tech startups in comparison to the corporations that already dominate the industry.
The second step is to identify venues or settings for the learning experience, including through outreach to instructors or student interest groups, or through workshops offered through the library. Ideally, these competencies can be integrated across the JD program by identifying specific opportunities where they make the most sense (see Table 3 below). And these competencies fall fairly logically along the three-year course of a JD program. The “know” competencies can be targeted more easily in an introductory first-year course, while upper-year courses can target the competencies associated with “access,” “evaluate,” and “use,” by introducing these concepts in a tangible way alongside the use of specific tools or in advanced legal research courses. The “ethics/legality” competencies naturally make the most sense to place in a legal ethics or professionalism course.
To give an idea of how these skills could be built across these different courses, consider the following example using the curriculum at Queen’s Law. First-year courses include an introductory legal skills course as well as courses in core areas of law, such as public law, contracts, criminal law, and torts. In keeping with the “Know” competencies outlined below, a very brief introduction to legal technology and AI could be incorporated into the legal skills course alongside a discussion of the duty of technological competence. Subject-specific tools that allow students to identify how AI is used in different areas of law could then be introduced in those other core courses—for example, the role of AI-assisted contract drafting and review tools in Contracts, the effects of judicial analytics tools in Public Law, and the use of problematic AI-driven tools in sentencing in Criminal Law. These are all high-level examples of raising a basic sense of awareness about AI in law for students in these first-year courses without delving too far into the use of specific tools or strategies for integrating them into practice.
Upper year courses, in contrast, can focus more strongly on integrating these tools and best practices into the curriculum. At Queen’s Law, students are required to write a substantial term paper in an upper year course, which provides an ideal opportunity to integrate information literacy competencies associated with many standard legal research tool, such as an understanding of how algorithms affect the outcomes of their searches and how to effectively use and evaluate the results of an AI-driven tool in the course of legal research.
Other substantive courses could also integrate more hands-on use of specific AI-driven tools in accordance with these competencies. For example, due diligence tools (in a Corporate and Commercial Law course) and outcome prediction tools (e.g., in Family Law, Labor and Employment Law, or Criminal Law courses).
Lastly, higher level competencies could also be fostered in experiential learning settings within the law school. For instance, training on AI-driven citation assistance tools may be provided to student law review editors, legal research tools to mooting teams, and practice-specific technologies to students working at law clinics.
By looking at AI through an information literacy lens, it becomes possible to identify specific competencies that should be fostered throughout a law school curriculum. This method can be helpful for designing instruction on this topic using the principles of backwards design and also for determining where specific competencies should be fostered at different points in a law student’s educational development.
In my own experience, student interest in this subject is high, and as a result, they are likely to be highly engaged in the classroom. Additional conversations on the integration of AI-driven tools into legal research and writing instruction would be fruitful to help instructors develop such sessions.
See, e.g., Emily Janoski-Haehlen & Sarah Starnes, The Ghost in the Machine: Artificial Intelligence in Law Schools, 58 Duq. L. Rev. 3, 22-23 (2020); Iantha M. Haight, Digital Natives, Techno-Transplants: Framing Minimum Technology Standards for Law School Graduates, 44 J. Leg. Prof. 175, 178 (2020); Melanie Reid, A Call to Arms: Why and How Lawyers and Law Schools Should Embrace Artificial Intelligence, 50 U. Tol. L. Rev. 477, 482-88 (2019); Emily Janoski-Haehlen, Robots, Blockchain, ESI, Oh My!: Why Law Schools Are (or Should Be) Teaching Legal Technology, 38 Legal References Service Q. 3 77, 83-88 (2019).
Model Rules of Pro. Conduct r. 1.1 cmt.  (Am. Bar Ass’n 2020); Model Code of Pro. Conduct r. 3.1-2 cmt. [4A], [4B] (Fed. of Law Soc. of Canada 2019).
Janoski-Haehlen & Starnes, supra note 1, at 98.
Am. Bar Ass’n, List of ABA-Approved Law Schools, https://www.americanbar.org/groups/legal_education/resources/aba_approved_law_schools/in_alphabetical_order/ [https://perma.cc/Y9BG-VQ85] (last visited Mar. 18, 2022).
See, e.g., Jamie J. Baker, Beyond the Information Age: The Duty of Technology Competence in the Algorithmic Society, 69 S.C. L. Rev. 557 (2018); Paul D. Callister, Law, Artificial Intelligence, and Natural Language Processing: A Funny Thing Happened on the Way to My Search Results, 112 Law Libr. J. 161 (2020); Susan Nevelow Mart, Every Algorithm Has a POV, 22 AALL Spectrum 40 (2017); Theresa Tarves, AI in Legal Education, in Law Librarianship in the Age of AI, 103 (Ellyssa Kroski ed., 2020); Haight, supra note 1.
For example, see the discussion of the term’s definitions in Woodrow Barfield, Towards a Law of Artificial Intelligence, in Research Handbook on the Law of Artificial Intelligence 2, 3-4 (Woodrow Barfield & Ugo Pagallo, eds., 2018).
For specific examples of tools, see Heidi W. Heller, Types of AI Tools in Law, in law librarianship in the Age of AI, supra note 5, at 31.
See, e.g., Susan Nevelow Mart, The Algorithm as a Human Artifact: Implications for Legal [Re]Search, 109 Law Libr. J. 387 (2017).
See, e.g., Margaret Martin Barry, Practice Ready: Are We There Yet?, 32 B.C. J. L. & Soc. Just. 247, 252 (2012).
Ellie Margolis & Kristen Murray, Using Information Literacy to Prepare Practice-Ready Graduates, 39 U. Haw. L. Rev. 1, 23 (2016).
Ass’n of Coll. & Res. Libr., Framework for Information Literacy for Higher Education 8 (Jan. 11, 2016), https://www.ala.org/acrl/standards/ilframework [https://perma.cc/XQ65-LXJZ].
For an idea of how this concept has transcended the complex technological advances of the past 30 years, note how similar these competencies are to the original proposal of information literacy as published in 1989: Am. Libr. Ass’n, Presidential Committee on Information Literacy: Final Report (Jan. 1989), https://www.ala.org/acrl/publications/whitepapers/presidential [https://perma.cc/6X3S-GPLY].
Am. Libr. Ass’n, Information Literacy Competency Standards for Higher Education 2-3 (Jan. 18, 2000), https://alair.ala.org/bitstream/handle/11213/7668/ACRL Information Literacy Competency Standards for Higher Education.pdf?sequence=1 [https://perma.cc/QG22-W6HG]. These competencies were later reorganized into the Framework for Information Literacy for Higher Education, supra note 11, which is the current iteration of the document.
Am. Ass’n of Law Libr., AALL Principles & Standards for Legal Research Competency (Apr. 2020), https://www.aallnet.org/advocacy/legal-research-competency/principles-and-standards-for-legal-research-competency [https://perma.cc/9JQ4-N4WL].
See, e.g., Casandra M. Laskowski, AI Defined: Core Concepts Necessary for the Savvy Law Librarian, in law librarianship in the Age of AI, supra note 5, at 1; Michael Legg & Felicity Bell, Artificial Intelligence and the Legal Profession: Becoming the AI-Enhanced Lawyer, 38(2) U. Tas. L. Rev. 34 (2020).
This is not a comprehensive adaptation of the framework, and there are likely further competencies that could be considered here.
“Backward design” is a framework first proposed by Grant Wiggins and Jay McTighe as a method of curriculum design that starts with identifying the desired results of a course or lesson, and then works backwards to plan the educational experience around those results. See Grant Wiggins & Jay McTighe, Understanding By Design (2nd ed. 2005). For examples of backward design in legal education, see, e.g., Nancy B. Talley, Are You Doing It Backward? Improving Information Literacy Instruction Using the AALL Principles and Standards for Legal Research Competency, Taxonomies, and Backward Design, 106 Law Libr. J. 47 (2014); Susan Azyndar, Work with Me Here: Collaborative Learning in the Legal Research Classroom, 1 Legal Info. Rev. 1 (2015-16); Gregory M. Duhl, Equipping Our Lawyers: Mitchell’s Outcomes-Based Approach to Legal Education, 38 Wm. Mitchell L. Rev. 906 (2012).
Ass’n of Coll. & Res. Libr., supra note 11, at 8.
See, e.g., Anna Yevelson-Shorsher & Jenny Bronstein, Three Perspectives on Information Literacy in Academia: Talking to Librarians, Faculty, and Students, 79(4) Coll. & Res. Libr. 535 (2018).
For example, librarians supporting law schools in northern Canada have commented on how barriers to internet access prevent students from accessing basic legal research tools like Westlaw and Lexis Advance Quicklaw. E.g., Aimee Ellis, Emily Tsui, Serena Ableson & Greg Hughes, Legal Information from Canadian Territories at the Canadian Association of Law Libraries 2021 Virtual Conference (Jun. 4, 2021).
Queen’s University, Juris Doctor (JD) Program, https://www.queensu.ca/academic-calendar/law/degree-programs/jd/ [https://perma.cc/MZ24-QY7K] (last visited Mar. 18, 2022).