Generative AI In Teaching

I teach college for a living in the Legal Studies department at a community college in Maryland. I have developed this site for my colleagues that are interested in integrating generative Artificial Intelligence (GenAI) into their teaching practice. What follows are some use cases for GenAI to help improve your teaching, reduce time spent on administrative tasks, and unleash your creativity.

In my opinion, GenAI can be used to improve your teaching by helping you focus on the best part of what you do. GenAI is not a replacement for faculty, but a force multiplier.

I’d like to start with your roadmap: more specifically, your plan for your course design. Generally, college courses have a list of the overall course objectives that describe what students will get out of taking the class. For example, one class I teach is Civil Litigation which has a total of nine course objectives. One of these is to be able to explain the major steps in a case from initial client interview through trial, judgment, appeal, and post judgment collection.

 If you've ever been involved in litigation, you know that there are a lot of steps between the initial dispute and a final judgment. That one course objective could require a fair amount of course time for a student to understand and be able to explain these procedures.

One theory of education is a constructivism, which posits that the learner is central to the creation of meaning. Biggs discusses this idea in his article, “Enhancing teaching through constructive alignment,” where he defines teaching as a system that embraces “teacher, students, the teaching context, student learning activities, and the outcome…” where curriculum and assessment methods are aligned.[i]

Garrison discusses the importance of aligning course objectives and the use of technological tools in support of those objectives.[ii]

For the instructional designer (including you, my dear reader), our design has to take into account how a student would achieve the objectives of the course. Given the breadth of the typical course objective, a wise instructor might define more specific learning objectives that would help the student attack the course objective in smaller parts. In my example, I would probably need more specific learning objectives that are related to individual steps in the litigation process.

Many of us teach using a textbook, and our textbook may have a chapter or two that relates to a particular course objective. For example, we use an Open Educational Resource (OER) textbook for Civil Litigation, and I have an eleven-page section of that textbook that discusses the general process for litigation disputes. Many textbooks may define specific learning objectives at the start of the chapter, and for convenience you may simply adopt or adapt these learning objectives for your course.

However, you may be using various resources, and not all may have pre-defined learning objectives. Or the ones provided in your textbook may not be what you want to focus on to achieve a specific course objective. So, one way to help you think about your learning objectives is to use GenAI.

My extract from the OER on the litigation process identifies the initial pleadings that the parties file at the beginning of the case: the complaint and the answer. It also talks about some of the other rules of civil procedure that govern what's supposed to be in that particular document, what’s required to demand a jury trial and so on.  But there are no defined learning objectives.

Let’s ask GenAI for help! With a number of the AI’s available today, such as ChatGPT and Claude.ai, you can upload documents to the system. In my case, I uploaded my textbook extract, and then I crafted a prompt: “Hi, Chat. Can you use this course objective ‘explain the major steps in a case from initial client interview through trial, judgment, appeal, and post-judgment/collection’ and the enclosed document to craft some learning objectives for my students.”

Chat will then respond with some proposed learning objectives based on your prompt and the attached document. Now, you might not like its initial ideas, but they can be a starting point for your own thoughts here. Think of GenAI as a fellow course designer that you are working with to solicit feedback and brainstorm with. In my case, Chat had ten ideas in total, including: “Describe the components of a complaint and explain its role in initiating a civil case” and “identify the requirements and methods for service of process.” Both of these are serviceable ideas to introduce students to how litigation generally is initiated, particularly for paralegal students who are likely to be called upon in a litigation firm to prepare these types of documents and file them with the appropriate entities.

Bloom’s Taxonomy

Our course objective asks students to “explain” the major steps in litigation. Before one could “explain” something, some prior learning would be needed. For example, I can’t explain concepts to you that I can’t define. Part of my explanation may also involve analogies or applications of these definitions to particular circumstances. The implication is that there is probably a definable learning process with a particular starting and ending point. Queue Bloom’s!

Bloom’s Taxonomy[iii] is a classification system that is designed to categorize learning objectives into categories, so that objectives can be organized from simple to complex learning tasks. One might think of organizing student learning into “crawl, walk, run” activities, where a learner must first know how to crawl before they can walk, and walk before they can run. The Taxonomy, revised by Anderson & Krathwohl in 2001,[iv] uses the verbs of a course objective – what the student is expected to do in the course – to categorize the objective. Lower-level objectives involve identifying, defining or recalling information, where higher level objectives might involve the application of knowledge, contrasting different ideas, or creating new material from knowledge gained in the course.

If you were to examine my learning objectives against Bloom’s Taxonomy, you would likely conclude that “describing” and “identifying” something is level one or two in the taxonomy – where students are first crawling with the concepts in the course. Given that we are introducing students to a complex idea – the process of litigation – one likely effective way to achieve this goal is to start at the start.[v] What is a “complaint” and what goes in it? What is “service of process” and what are its requirements for a defendant to be properly served the complaint? Chat’s suggestions for these two learning objectives seem like a good place to start.

A course design should also include the instructional materials for the student. In my course, that would include the OER textbook itself and probably the applicable Rules of Civil Procedure (we use the Maryland Rules). You may need to consider what else a student would need to get started working towards the learning objectives you have defined as you develop your course design.

The next question for your design is what students will do with these materials to achieve the learning and course objective. Our agenda should be to align these activities with our course and learning objectives. [Discuss Garrison]

Learning Activities

Learning activities should be aligned with the learning objectives to help support your students. So, then, what should we have our students do? Obviously if you provide them with reading materials, they should read them. But what then?

Well, you could ask your design partner for some ideas. Your prompt to Chat could be along the lines of: “Can you develop some ideas for students to achieve the first two learning objectives you have listed above, after they have read the textbook?”

And Chat has a lot of ideas. In response to the above prompt, it produced five in total, two for the first learning objective on the pleadings, and three on service of process.

One of its ideas is to give the students a sample complaint and ask them to identify its parts, such as the jurisdictional statements of the parties, the material facts of the case, causes of action and requests for relief from the court. This could be an individual or small group activity for students.

Another idea Chat suggested was to have students draft a complaint based on a scenario using a template. For example, in Maryland, the district court is the limited jurisdictional court. The court website provides several forms that litigants may use such as a form for a Complaint.

This kind of activity can also be impactful on student learning, but this activity is probably a higher-level one as students are applying (or for a general jurisdiction complaint in our circuit court, creating a document using) information acquired earlier in the class. Such an assignment may fit later into our overall design.

Ideas are great, but for successful instruction, you must develop the exercise so that your students can successfully complete it. We could ask Chat to develop a complaint identification activity using learning objective one by crafting a prompt like: “can you develop a complaint for use in the activity 1 that you list for learning objective 1 above?”

For me, it generated a basic complaint of Mary Johnson v. XYZ Painting LLC with various components of a contract dispute. I can copy this into a Word document and edit it to develop a final assignment that I can give to my students to identify the parts of the complaint correctly. Such an assignment would help align an activity in my class with a learning objective.

You should note, however, that Chat isn’t designing the whole course autonomously. Our prompts are breaking down elements of the course design into concrete, actionable responses from the GenAI over which we ultimately must exercise discretion as to how or whether a suggestion fits our overall design. GenAI does not supplant but instead enhances our capacity to create better courses.

Assessment

Another issue to consider is how I am going to assess student learning of the learning objectives that I've set out for them. A key consideration is how the assessment aligns with my course & learning objectives and the course assignments.

As noted above, my learning objectives in this example are relatively low-level Bloom’s-type objectives. In planning for assessment, I would want to consider the objective so assessed to align  my assessment. Therefore, an appropriate assessment might be for students to correctly answer several multiple choice items concerning the elements of a complaint or the steps for properly serving a defendant.

Unsurprisingly, Chat could also help you with developing an assessment. Your prompt might look like: “can you generate some multiple choice questions related to the first learning objective we defined above?”

Again, you may not like every question it generates, but it gives you some ideas to start. This is particularly impactful if you don’t have publisher materials like an instructor’s manual to create these items.

This is also another way to brainstorm with GenAI to generate some items that are aligned with the course materials and learning objectives. You can also ask Chat to adjust the learning level of the questions generated to match the level of the learning objective to really fine tune your assessments to the other parts of your roadmap.

Now this is a lot of work to do by hand, which is one of the reasons why artificial intelligence can be so impactful. If I'm teaching a new class, it may take me most of the semester, realistically, to build the whole course out, from reading the textbook and developing materials and everything else by hand: that's a lot of work to do that.  GenAI has the capacity to really compress the time it takes to develop a course design. Again, not every idea it has is great. Not every idea is aligned. Not every idea is going to work. But GenAI can generate ideas and examples of activities that you can use, like a cookbook, which you can use to further innovate.