A tool built in an LLM positioned 50 modules of content across a graduate curriculum.
Dr. Manuel Ruiz-Aravena, an assistant professor of wildlife disease ecology and One Health in the Forest and Wildlife Research Center (FWRC), is building a curriculum analysis tool in Claude to take on the mechanical parts of curriculum design.
By AI Hub Team
Ruiz-Aravena’s interest in AI for curriculum work started with an NSF project. He collaborated with partners at Bentley University and Montana State to develop a curriculum examining the complex relationships among environmental, human, and animal health, economics, policy, and land use. During that work he gained an appreciation for how well Large Language Models (LLMs) could navigate the varied factors represented in One Health projects and assist in developing a curriculum that explores the intersectionality of broad topics.
He is now applying that understanding with FWRC Associate Teaching Professor Dr. Leslie Burger as the two of them reassess the Master of Conservation Education curriculum. Ruiz-Aravena explains that analyzing a curriculum is a complex venture. Courses are vertical, and students progress in time through courses that are connected to programs. The mapping is a three-dimensional process, and it is usually limited by the number of hours a person can give it. Ruiz-Aravena’s goal is to offload the mechanics of that mapping to an AI-assisted tool. Using Claude, he has developed a set of instructions and skills to build a curriculum analysis tool to assist in the task.
Burger began by identifying the competencies specified by the program’s accreditation organization and mapping those competencies across all six core courses. Ruiz-Aravena’s instructions to Claude included Burger’s mapping, details on the cognitive level at which each competency should be taught, and direction on how the knowledge should scale across courses to meet program objectives. He then processed Burger’s existing course content to make it structured for machine reading. The LLM then systematically processed that information to provide structured outputs and position 50 modules of content across the curriculum, along with developing plans for slides and transcripts that instructors can review and adjust as needed. Everything that the system assesses is based on instructor definitions and content.
The system also allows flexibility. When instructors want to bring in additional papers and examples, the LLM can be directed to include these additions and adjusts the scripts and slides accordingly, while also keeping in sight the impact of changes in the course and program sequence. This is the part that excites Burger. “This tool takes care of the time-consuming parts of curriculum design, allowing the instructor to focus on the creative process in impacting students,” she says.
Ruiz-Aravena points out multiple uses for the tool. A curriculum committee can use it to analyze the coherence and sequence of courses and to determine where redundancies and gaps in content might occur. The committee provides the LLM with course materials and instruction directing specific workflows to review a course in the context of the broader program. That step offloads the mechanical and normally time-consuming part of the curriculum review. Committee members, as subject matter experts, are then able to invest their time in giving faculty personalized direction on any needed curriculum, assignment, or assessment adjustments in a course or course sequence.
The tool can also be helpful to individual professors, who can use it to check that syllabi, assignments, and assessments accurately reflect the stated learning objectives and competencies of a course. The LLM can help evaluate whether the material is being taught and assessed at the appropriate cognitive level of Bloom’s Taxonomy and whether the advanced courses in a program reflect higher cognitive levels. This is a particularly useful check for faculty who are not fully integrated in all the courses in a sequence.
In each of these uses, the judgment stays with the faculty. The LLM assists with the mapping and produces the drafts, and the instructors and committee members review that output and decide what is accurate and what belongs in the course.
Ruiz-Aravena is enthusiastic about what the tool could mean for his department and for his own instructional practices, and he looks forward to sharing his experiences with others who would like to know more.