Assignments have to fit into our learning objectives, be related to the course content, and challenge students to really master what we''re trying to teach. For those of us who teach in the legal field—particularly for paralegals—the goal is to prepare students for practice. That means thinking about what a paralegal would actually do in real life, and what assignments would reflect that.
For example, being able to conduct legal research and find relevant cases and statutes is a practical skill. Drafting documents is another. In a litigation firm, there are many documents that need to be created—from initial pleadings to discovery requests to motions and memoranda. That kind of work is directly related to what a paralegal might do, and it''s something we can simulate in class.
I teach a number of classes, including a course on business law specifically for paralegals. I approach this course as transactional training—meaning we focus on document generation rather than litigation. You could certainly teach business law as a litigation course, but I choose to focus on what a paralegal in a transactional business law firm would actually do. That leads to the question: What assignments can I give that are demonstrative for students?
What I''m sharing here is a sample assignment based on an operating agreement from Nolo and a copy of the statutes governing limited liability companies (LLCs) in Maryland, found in Title 4A of the Corporations and Associations Article. The agreement I have is not complete—it''s just the first few pages—but it''s a good starting point. The scenario is that we''re going to form an LLC for a client. Besides the Articles of Organization, the other typical document generated is some form of operating agreement. A paralegal might reasonably be expected to draft the first version of that.
This assignment requires some foundational instruction: What is an LLC? Who owns it? How is it managed and funded? These issues would be covered in the course. But ultimately, the output would be the document—something that an attorney would review before it goes to the client.
Now, the dirty little secret of law firms is that we tend to reuse the same forms—at least for initial drafts. If you''re in a business law firm, you probably already have a form operating agreement. It makes sense to start with that and tailor it to the client''s needs. But what if the form is outdated? Maybe the law has changed. The legislature updates the statutes every year. Does your form still comply?
So for this assignment, I plan to have students use AI to help with the analysis. I''ll provide ChatGPT with both the statute and the operating agreement (with the blanks filled in). Then, I''ll ask the students to map the provisions of the agreement to the relevant statutes in the Corporations and Associations Article.
Typing this all out can be a bit of a chore—I've done it before—but the idea is to experiment. Students will probably get different responses from the system. That''s fine. The purpose is to give them a starting point—a five-page excerpt from the agreement—and ask: What matches what? Which provisions map to which statutes? And if they don''t align, is there a reason?
For example, having an effective date might be practical but not legally required. Other provisions are more clearly dictated by statute: formation requirements, naming conventions (under §4A-208, which refers to another section in Title 1), business designators in the name, having a resident agent, a principal office in Maryland, a stated business purpose—these are all required elements.
This goes on to include sections about tax and financial classifications. In Part Three of the agreement, there are six paragraphs—six different provisions—that can be matched to statute sections. That''s a good starting point for students. The assignment could then ask: What must you do to comply? Does the statute require something more? Are we missing information from the client''s fact pattern that would help us answer these questions?
This becomes a compliance tool—or at least a starting point for ensuring compliance between a draft document and the governing statute. Manually comparing 150 pages of statutes with a 12- to 32-page agreement is a heavy lift for a student—or even for a lawyer. Demonstrating that this process can be partly automated is powerful. It helps students understand how practice might actually work.
This also gives us a chance to discuss the ethical boundaries of using ChatGPT. I''m only giving it a blank form—no client data. But if you feed the finished form into the system, you run the risk of a confidentiality breach. AI systems like ChatGPT don''t owe you confidentiality, and whatever you submit might be used to train the model. That''s partly why it only costs \$20 per month. So, before submitting a document, students should understand the risk. Using a template avoids those problems, but it''s also an opportunity to teach about legal ethics and confidentiality.
This could be expanded into an assignment where the student generates additional prompts. For example: “What''s missing from this operating agreement?” ChatGPT, having “read” the statute and the first five pages of the agreement, will probably come up with quite a list. That''s useful. It prompts students to ask: Do we have a complete agreement?
Some of the work here is dependent on the fact pattern. What did the client want? Who are the members? Who''s the manager? What are the capital contributions? We likely need to address things like the inspection of records, the process for winding up the LLC, how to admit new members, transfer membership interests, and what voting thresholds apply—majority, supermajority, unanimous?
These are all good issues to raise. Students could then be asked to draft those missing paragraphs, or at least some of them, as part of the assignment. Ideally, I would have given them a more complete agreement, or we would have found a complete form to work from. But this still illustrates how ChatGPT can be used practically—by students and even by law firms—as a tool to identify gaps or to locate provisions in alternative forms.
The key point is that this integrates AI into the assignment without surrendering the student''s agency. You can''t just ask ChatGPT to do the whole assignment for you. You have to engage with the output—validate it, check the statute, conduct additional research, and apply it to your client''s scenario. That''s how you create a complete and usable work product.
This is a sophisticated assignment. In the past, when I asked students to do this manually, they might find a few relevant statutes. But here, they get a more comprehensive view. The next step is to act on that view—to create a full, usable document. That''s what law firms want: a full, mostly accurate draft that can be reviewed by the supervising attorney.
So again, this is a more sophisticated way of using AI in assignments—one that enhances, rather than replaces, student learning. We want students to gain something from the process, not just delegate it all to AI. Because AI isn''t always right, and it needs supervision.