Teaching Law Students to Write in the Age of AI

After GenAI use became widespread among students worldwide, reactive guidance for teaching and learning counseled avoiding written work as the basis for assessments – in other words, de-emphasizing writing because those assignments are most susceptible to undetected unauthorized GenAI use. While this approach may work well for small class sizes (where individualized, oral exams are possible, for example) and substantive courses (where learning specific content is the chief goal), this approach fails for writing courses for which the primary purpose is to determine whether students can successfully compose writing in different contexts, for different audiences.

As the course coordinator for the Legal Research and Writing (LRW) program in HKU’s Faculty of Law, each fall semester I teach between 260-300 first-year students the crucial components of legal writing for eventual practice. In LRW, students are expected to demonstrate their legal reasoning skills, to write succinctly and effectively, and to construct logical and persuasive arguments that reflect understanding of and ability to adapt to a particular situation or client scenario.

Unable to abandon writing as a skill or ignore the reality of GenAI, over the past 3 years the LRW curriculum has adjusted to embrace AI, teach thoughtful and ethical engagement with the multiple new writing tools available to students, but also to show where these tools currently fall short or where their outputs must be treated with skepticism and thoughtful caution. It is a delicate balance to strike, one made more difficult by several external factors, including the opacity of reliance on student acknowledgment, rapidly improving quality of AI outputs, and the unreliability of AI detection tools.

It has been important to underscore for students the rationale for written tasks (especially skill development), emphasize the reasons to trace information back to primary sources, and to stress students’ personal responsibility for the accuracy of anything submitted, whether using GenAI or not.

For LRW, effective activities in the classroom include using GenAI explicitly in practice exercises, sharing prompts / outputs with students as part of the learning process, and encouraging students to critique the substance and style of GenAI writing in response to legal scenarios. Multiple exercises designed to guide students to use GenAI at different stages of writing / editing allow students to experiment and to see many examples of the potential benefits and limitations of GenAI as a writing tool.