From Builders to Editors: Reimagining How We Teach Citation in the AI Era

Introduction

For generations, law students have been taught to treat citation as a construction project. They were expected to memorize rules, assemble components—signal, case name, reporter, pinpoint citation, date—and produce a finished product that matched the exacting format of The Bluebook, ALWD Guide to Legal Citation, or another legal citation manual.^[1] This “building” model was justified when lawyers were likely to draft most of their citations from scratch. But that reality has been fading for a number of years and is no longer a realistic approach in the age of generative artificial intelligence technology.

I. The Shift Away from the Building Model

When preparing to teach citations in a legal writing class, professors must first choose the source they will rely on for their teaching, such as The Bluebook or ALWD Guide, design and create study questions and exercises, and supplement that material with online resources and testing.^[2] That method focuses on building citations from scratch.

However, in practice, because lawyers are expected to complete their tasks as quickly as possible, lawyers have long relied on recycled citations gained from other sources. These recycled citations could come from copying from prior briefs, borrowing from templates, or lifting ready-made citations from Westlaw, Lexis, or Fastcase.^[3] Now, with the rise of generative AI, the shift is even more dramatic. Lawyers and students can generate seemingly polished citations in seconds.^[4]

The problem? These machine-produced outputs can be incomplete, imprecise, or entirely fabricated.^[5] Lawyers remain responsible for what goes into the final document, but the process of how they get there is no longer a linear process that relies solely on the expectation that they will “build” the components into a citation generated solely by their own brain and knowledge of citation rules.

Since the widespread adaptation and integration of generative artificial intelligence technology, the way members of the legal profession draft and use citations has shifted even further away from the traditional “citation building” model. This shift has created reliability problems and new potential for sanctions and humiliations, as lawyers have been caught submitting briefs to courts that cite cases that do not actually exist.^[6] Recent research shows that OpenAI, Anthropic, Google Meta, and Deep Seek already make active attempts to generate fully compliant Bluebook citations.^[7] As of early 2025, these attempts are only accurate 69% to 74% of the time, and even when these artificial intelligence platforms were trained with in-context learning of The Bluebook’s rules, that training only raised accuracy to 77%.^[8] But the temptation to save time with generative AI will only increase as AI gets better at mimicking citation formatting rules^[9] and as pressure from clients for lawyers to work faster and more efficiently continues to build.^[10]

In this rapidly changing landscape, law professors now face increased pressure to prepare students for a new way of working.^[11] To better prepare students for the modern process of legal writing in practice, legal writing professors should start considering how they can change the focus of their teaching citations to focus more on editing citations procured from other sources instead of drafting them from scratch.

II. Implications for Teaching Citation

With that shifting curricular focus in mind, now is the time for law professors to find new ways to teach the language of legal citations. Citations are a critical component of ethical and reliable legal writing because citations are the backbone of the language that lawyers use to communicate with each other and the court about the law.^[12] Understanding the reading and drafting of legal citations is a vital legal skill.^[13] The pedagogical approach that law professors use to teach students how to read and draft legal citations is crucial^[14] because students need this skill to be effective and ethical advocates for their future clients.

In the next few years, many law professors, especially legal writing and clinical professors, will develop innovative instructional strategies that embrace new citation generation technologies. When considering options for new instructional strategies, law professors should bear in mind several pragmatic considerations. The reality is that students increasingly procure “complete” citations from online sources and legal databases.^[15] Generative AI platforms will improve in their ability to learn and replicate the picky formatting rules of complicated citations, further encouraging law students and legal professionals to use generative AI for procuring complete citations. If current trends continue, then it is only a matter of time before most legal professionals use completely AI-generated citations in their legal writing.^[16] In this new environment, teaching students to ethically evaluate a citation for reliability will be even more vital than teaching them to generate a correctly formatted legal citation from scratch.^[17] Perhaps most importantly, teaching students to edit an existing citation for accuracy will likely be a more useful skill than teaching students to “build” a new citation from scratch.^[18]

This evolving reality is increasingly at odds with the underpinnings of widespread methods for teaching legal citation skills. The traditional pedagogical approach treats citation generation as a building exercise, in which students craft citations from scratch. Whether professors use lectures combined with worksheets and quizzes, exercises from online platforms, or interactive books from legal education sources, the traditional focus is usually on identifying each part of the citation, element by element, to teach students how to “build” their citation sentence by locating and replicating the citation rules for signals, case names, traditional reporters or online sources, pincites, court information and dates, and typeface. Online lessons using products like Lexis Interactive Citation Workstation^[19] or Introduction to Basic Legal Citations through CALI Lessons also focus their instruction on these individual elements for the purposes of replicating formats so students can “build” their citations, element by element, from scratch. While some professors use gamifying tools or classroom response technologies such as iClicker,^[20] even these more recently developed methods assume that students need to know how to create their own citations from scratch.

However, that assumption may be largely false. Not only are students and lawyers increasingly using citations previously generated elsewhere through simple “copy and paste” functions, but lawyers have already done so in practice for many years.^[21] Whether lawyers are using a previously drafted memo, motion or brief as a starting point for their own memo, motion or brief, or whether lawyers copy and paste citations from Westlaw or Lexis, the fact remains that lawyers have been using previously generated citations in practice well before the release of ChatGPT in 2022.^[22] As such, perhaps the focus on teaching students how to draft legal citations from scratch has been misguided for quite some time. While law students need a solid understanding of the components of legal citations to effectively edit citations procured from other sources, the reality is that citation rules and legal technology will continue to evolve at a rapid pace. Just as in other aspects of the law, the ability to locate and reference the current rule is going to be more important to future lawyers than the ability to execute the rule as it existed when they were law students. Thus, the ability to execute editing skills, and keep those editing skills up to date with the latest technology, is the more valuable skill for future legal professionals.

No generation of lawyers has ever been judged on how quickly they can recall Table 6 abbreviations or format small caps. Instead, lawyers will increasingly be measured by their ability to:

• Critically evaluate authority. Can they determine whether a precedent exists? Whether it truly supports the proposition? Whether it is still good law?

• Verify AI-generated outputs. Can they recognize hallucinations, omissions, or subtle misquotes from automated systems?

• Balance speed with responsibility. Can they work efficiently without sacrificing ethical standards?

• Adapt to evolving tools and norms. Can they approach new research technologies with both curiosity and healthy skepticism?

In other words, the future of citation literacy lies less in assembling citations and more in curating them—ensuring accuracy, transparency, and integrity in the final product.

To evolve teaching methods to meet the challenges, opportunities, and pragmatic realities of the age of legal citations generated by any AI platform, law professors should consider changing their focus when teaching citations from a building perspective to an editing and ethics perspective. In this new approach, the educational objective is not to teach students to build citations from scratch, but instead to teach students how to critically evaluate previously generated citations to ensure they are accurate, reliable, and ethically compliant.^[23] To meet these three objectives, the starting pedagogical premise must focus on teaching students how to first ensure the reliability of a citation before reviewing it for technical formatting issues.^[24]

By shifting the emphasis to ensuring the reliability of the citations, law professors can help students prepare for the future by focusing less on the technical details that will eventually be outsourced to computer programs and artificial intelligence. Instead, valuable classroom time and educational outcomes can be refocused towards ensuring that students understand how to test and edit the accuracy of the information contained in a previously generated citation.

Eventually, lawyers may not need to focus on the detailed nuances of citation typeface and formatting.^[25] However, given the rules governing lawyers’ professional responsibility and ethics, what will never change is the need for lawyers to remain focused on the ethics and reliability of the legal citation they use to support their legal arguments. And while the traditional rules of typeface and spacing should not be completely abandoned during the teaching of legal citations, given lawyers’ increased use of previously generated citations in the practice of law, more time must be spent helping students develop an understanding of the need to balance the ethical ideals of citation accuracy with the realistic practice pressures related to time and accuracy.

III. Steps for Switching the Focus from Building to Editing

To adjust to a new and evolving reality, legal writing professors will need to make structural and practical changes to how they teach citation. Specifically, legal writing professors should bring AI into the classroom and shift the focus to teaching students how they can ensure the reliability of citations.

A. Bring AI into the Classroom

In general, future legal citation lessons should be designed to focus on editing skills. While law professors may continue using traditional methods of teaching legal citations and assessing student competencies in citation skills, professors should also start allowing students to use AI-generated citations so students can get practice with these technological advances before entering the practice of law.^[26] One option is to generate practice exercises using currently available AI platforms such as Microsoft Co-Pilot, Claude, or other available AI generation models. These lessons should include warnings about the dangerous potential for artificial intelligence hallucinations and incorporate strategies for how to catch inaccuracies and hallucinations, along with methods for conducting research in a continually evolving technical landscape.^[27]

For example, professors could consider:

• Including exercises or examples that require students to start with “completed” citations and teach them to catch reliability errors as well as typography errors;^[28]

• Including methods for evaluating citation reliability; and

• Being transparent about the technologies currently available to generate citations.^[29]

Such an approach provides “real world” examples of ways to test the reliability of a source or technology. Ideally, this approach also reinforces the understanding that technology will continue to evolve and change during students’ lifetimes. As a result, students’ learning objectives should include an understanding that they are expected to continually evolve as well.

Professors may also want to consider being transparent and authentic with their students about their own need to evolve. Professors have the potential to make learning about AI a partnership with their students by encouraging students to speak openly with professors about questions they may have, confusing AI encounters and outputs, or new products released to students or that students find independently related to AI use. Because AI is evolving faster than human comprehension of the technology and its capabilities, this type of open partnership has the potential to create opportunities for professors and students to learn and evolve together. Additionally, professors engaged in scholarship may want to experiment with using AI to generate citation formats for their own academic projects. While current popular academically-focused citation generation products include Zotero, Mendeley, and Juris M, more products will likely be developed in the future.

B. Teach Students how to Test Citations for Trustworthiness

How can law professors reorient their pedagogy to embrace the latest technological advancements? First, the goal should be reframed. Instead of teaching students to build citations piece by piece, law professors should be training students to test citations for trustworthiness. Doing so will automatically shift from a construction mindset to an editing and ethics mindset.

Practical classroom approaches might include:

• Addressing evolving practices. Discuss how lawyers generate legal citations in the “real world.”

• Starting with flawed citations. Give students pre-generated citations—some from AI, some from online databases—and have them identify and correct errors. This mirrors what they will do in practice.

• Integrating technology transparently. Show students how AI tools generate citations, then demonstrate where those outputs fail. Let students experiment, critique, and learn the limits of these tools under supervision.

• Teaching reliability first, formatting second. Begin by asking, Does the case exist? Is it accurately represented? Is it findable? Only after students confirm reliability should they refine typography and punctuation.

• Discussing ethics explicitly. Use classroom hypotheticals to explore real-world consequences of unverified citations—sanctions, loss of client trust, or judicial skepticism.

• Modeling lifelong adaptability. Encourage students to view citation practices as evolving, not static. The rules may change, but the responsibility for accuracy endures.

By embedding these methods into class exercises, law professors can help students internalize the idea that technology is a tool—not a substitute for professional judgment.

Conclusion

If we continue to teach citation as if students will always be “builders,” we are preparing them for a world that no longer exists. What they need instead is the ability to evaluate, edit, and ethically verify citations in an environment where automation is ubiquitous. By shifting their classroom focus to editing, ethics, and reliability when teaching legal citations, law professors can help prepare students to navigate the challenges and opportunities of generative artificial intelligence while preserving the foundational purposes of legal citations.

Teaching students to critically evaluate AI-generated citations equips students with the skills to ensure the accuracy of those legal citations, uphold ethical standards governing all practicing attorneys, and evolve alongside technological advancements—all of which are key competencies for success in an ever-changing legal practice landscape.