Designing a Course to Support P12 Educators in Using GenAI: A Design Case Informed by ChatGPT

This design case aims to share precedent knowledge and experience from designing a graduate course that empowers P12 educators to use generative artificial intelligence (GenAI) in their teaching context (Boling, 2010; Howard, 2011). Through thick description, this case aims to share the process, decision-making, and critical reflections of designing the course, including using ChatGPT to create instructional materials and provide feedback on the course design and development. This case includes the challenges we encountered and precedent knowledge for designers (Smith, 2010).

After the release of GenAI chatbots, many tools entered the ed-tech market, promising to support educators (Trust et al., 2023). Some teachers have embraced and use these tools daily (Lee & Kwon, 2024). However, many educators are unprepared to leverage these tools, and some remain hesitant to engage with GenAI tools because of ethical concerns (Trust et al., 2023; US Department of Education, 2024). These circumstances highlight the need for comprehensive support systems to assist educators in effectively using GenAI tools to improve teaching practices by designing courses that equip educators to use these platforms in the classroom (Seufert et al., 2021; Tate et al., 2023).

This design case highlights our journey of designing a course to support graduate students who are in-service P12 teachers using GenAI in the classroom and empower them to be leaders for GenAI adoption in their context. We used ChatGPT to support our design and development of instructional materials. This case aims to share the experience working with ChatGPT throughout the instructional design process and how ChatGPT-generated outputs influenced our design decisions. We aim to share design precedent and experience and do not seek to prescribe guidance or lessons learned (Howard, 2011). The following section will highlight the context of the design.

Context

Setting

This course was designed within an Instructional Technology and Design program at a comprehensive university in the southeastern United States. Students enrolled in our graduate programs can take this elective course in various program levels (i.e., master's or education specialist) and with varying concentrations (i.e., instructional technology or school librarianship). Each of these concentrations represents the largest programs in the state, with a collective of 1100+ graduate students enrolled in these programs. Approximately 98% of these students are currently practitioners in a P12 environment.

Delivery Model and Design Intentions

To best support our graduate program, which serves predominantly in-service teachers, the design team focused on developing a course that could serve as a traditional graduate-level course and four self-paced professional development modules to support teachers and school administrators in leveraging GenAI in their specific context. The intention was to provide opportunities for school districts to access our institution's high-quality, asynchronous professional development resources to support teachers in navigating GenAI in their context, which could be offered at scale. We still believe the course could serve as four separate professional development modules. However, as we continued to develop the traditional graduate-level course, this was not the main priority. As we developed the initial maps into the final course, we integrated more activities and authentic assessments that required real-time expert review and feedback. We intend to refine and iterate each module to be available for our school district partners to purchase as discrete, self-paced professional development modules. The course is designed and coded as 100% asynchronous; however, there are ample opportunities for synchronous interaction, including virtual meetings.

Institutional Constraints

The design team developed this course against the backdrop of broader university-wide conversations focused on how faculty would utilize GenAI in their instructional practices. These conversations, which were highly contentious, focused on student usage of GenAI and official syllabus policies, with policies for faculty usage of GenAI for various tasks related to their performance to come later. University policy adopted a four-option approach to student usage of GenAI, ranging from no usage to expected and direct usage within coursework, leaving the decision in the faculty's hands.

The course designers have broad P12 experience and have served in higher education for many years. However, another constraint of designing this course is that the designers have not previously used GenAI with P12 students. This constraint encapsulates the challenge of bridging research to practice and continuing to maintain relevance in an ever-changing instructional technology landscape. Therefore, this course for graduate students served as an investigation into understanding best practices and research surrounding GenAI in education.

Intended Learning Outcomes

Upon completing the course, learners will be able to take a more systematic approach towards GenAI for themselves and their organization, as they will learn how to use their knowledge to enhance their job performance in various ways (e.g., instructional and non-instructional tasks) and integrate tools and concepts related to GenAI into learning by identifying strategies and approaches to equip their learners to use GenAI tools. Furthermore, they will be exposed to the ethical dilemmas surrounding GenAI use, explore ways to diffuse it throughout complex organizations through careful considerations of appropriate policies for P12 schools, and identify effective GenAI professional learning methods. The following section will describe the members of the design team and the phases of the design.

Design Team and Phases

The design process occurred in two phases. In Phase 1, the design team comprised of six faculty members. The initial design lead for the course was an associate professor and interim department chair, supported by an associate professor, two assistant professors, and two clinical assistant professors. All team members teach graduate courses in instructional design and technology, change management, human performance improvement, online learning, digital game-based learning, and multimedia design. This design team developed the course and module learning objectives, course maps, and the modular structure at a retreat during a spring semester.

Phase 2 began in the fall semester when the initial design lead created a new design team consisting of one of the original members, a clinical assistant professor who is the first author, and a newly hired assistant professor. The clinical assistant professor teaches courses in instructional multimedia design and assessing technology-enhanced instruction and has expertise in maker-based learning and instructional technology coaching. This faculty member designed Modules 1 and 3. The assistant professor also teaches courses in assessing technology-enhanced instruction and change management and has expertise in simulations and artificial intelligence. This faculty member designed Modules 2 and 4. Both faculty members collaborated on Module 5. This new design team was tasked with developing the final design and teaching the course in the upcoming semester. The design team in Phase 2 used the course maps and materials from Phase 1 as a starting point for building the modules, activities, and assignments. The course maps developed in Phase 1 were entered into ChatGPT, followed by the prompt, “I am creating a course entitled Generative AI in P12 Settings. Can you analyze and correct these four documents, combine them into one, and ensure that the learning objectives align with assignments and assessments?”.

As mentioned in the introduction, the design team used ChatGPT as a tool throughout the design process. The outputs from our prompts were routinely checked for accuracy and adherence to the original course learning outcomes. The rationale behind using ChatGPT was to use the course design as an opportunity to learn the affordances and constraints of using ChatGPT to develop instructional materials and identify how it could support our instructional technology graduate program. While every team member has hesitations about the ethical implications of GenAI, all members are cautiously optimistic about the potential for GenAI to support P12 teaching and learning. Each output from ChatGPT was rigorously analyzed and scrutinized to ensure alignment with the course and the program's goal. The narrative below describes the design decisions we made when developing each module.

Design Process

In Phase 1, a survey (See Appendix) was distributed to existing students to evaluate their understanding of GenAI and the extent to which they apply it in their professional work. The survey was voluntary and not tied to course assignments and included Likert scale, multiple-select, and open-ended response questions. The needs assessment of 98 graduate students highlighted the need for hands-on GenAI practice, ethical guidelines, and policies.

ChatGPT was used to conduct qualitative data analysis by entering student responses from the open-ended questions in the needs assessment into data analysis-focused prompts. After reviewing the outputs, the design team conducted a thematic analysis to verify the output of the ChatGPT-generated text (Miles et al., 2019). Themes included academic integrity, ethical use of GenAI, teacher preparedness and policy, the role of teachers in education with GenAI, educational equity and access, and the impact of GenAI on learning and development. After engaging in a rich discussion regarding the course structure and the overarching goal, the themes from the survey were used to guide a content analysis of open educational resources to create the initial course maps and design. We arrived at the following as the overarching goal of the course:

The goal for GenAI in P-12 Education is to equip educators with the skills and knowledge to effectively integrate artificial intelligence technologies into teaching and learning practices, enhancing the educational experience for preschool through 12th-grade students.

As previously mentioned, one of the challenges in our design was to conceptualize the course as a traditional 16-week graduate course and four discrete instructional courses that could be deployed and used by local school districts for professional development. The initial four-module sequence included the following topics:

1. Introduction to GenAI in P12

2. Teacher use of GenAI

3. Learner use of GenAI

4. GenAI Leader in P12

The four modules introduce educators to the applications, strategies, and ethical considerations of using Generative AI in P12 education. By the end of the course, participants will be able to incorporate GenAI tools to enhance learning, increase efficiency in their teaching practices, and become GenAI leaders in their schools. The course learning outcomes include:

1. Identify nascent GenAI tools and their applications in P12 education.

2. Analyze the benefits and challenges of integrating GenAI into classroom teaching and administrative practices.

3. Develop GenAI prompts to enhance instructional materials.

4. Create policies and guidelines for the ethical use of GenAI in education.

5. Lead professional development on the use of GenAI in educational settings.

As the design team deliberated, we continually returned to our programs’ purpose (i.e., creating instructional design and technology leaders within organizations). Initially, module four, GenAI Leader in P12, was primarily focused on the ethical use of GenAI in P12 settings. However, one team member argued that when considering using GenAI with learners, an instructor would want to consider guardrails and that ethics and appropriate use of GenAI should be embedded throughout all four modules. Therefore, we modified the fourth module to emphasize becoming a leader in GenAI adoption and implementation within an organization. Thus, we moved forward with ethics being embedded into every module.

Once we finalized the overarching concept for the four modules, our team began building detailed course maps for each module. We envisioned each module taking approximately four weeks. Three designers each took the second, third, and fourth modules, respectively. The other two team members took the first. Our team divided and spent a day drafting the course maps. After creating these course maps, we debriefed as a group. Once the design lead approved the initial draft, we returned to our design partners and gathered other resources. At the end of the retreat, the course maps were complete with ideas for module learning outcomes, instructional materials, learning activities, assessments, and rubrics. After the retreat, the design team was committed to learning and effectively integrating GenAI into their teaching practices. Instructional materials such as videos and readings were frequently shared among the group throughout the spring and summer semesters.

In Phase 2 of the design, we recognized a deficit in our original design. We added a fifth module to encourage learners to reflect on their experience with GenAI in education through the course and provide feedback to improve future iterations. The final design consisted of five modules in which students examine the fundamentals of prompt engineering, utilize GenAI for administrative tasks and personalized learning, create a lesson that engages students with GenAI, and develop a plan for professional development to share their GenAI expertise with their colleagues. The following section will describe how each module was developed with the significant design decisions from Phases 1 and 2.

Module 1

Phase 1

In Phase 1, we envisioned Module 1 as an overview of GenAI in education and prompt engineering. In the course map, week 1 introduced artificial intelligence, and students were expected to share their opinions about GenAI in a discussion. During week 2, we intended for students to evaluate the affordances and constraints of GenAI tools in a discussion or analyze a case study. In week 3, we planned to introduce prompt engineering and ask students to share their process of creating a prompt in a discussion—lastly, week 4 concentrated on creating and refining prompts and analyzing the outputs in an assignment. An image of the course map for Module 1, Week 1 is shown Figure 1. The following sections will highlight the design decisions made throughout Phase 2.

Figure 1

Module 1 Week 1 Course Map

Phase 2

In Phase 2, week 1 was primarily left unchanged because it covered foundational GenAI concepts for the remainder of the course. When creating content for week 2, we realized the GenAI tool exploration discussion and case study would be more appropriate in Module 2 - Teacher use of GenAI. Around this time, we created a fifth module to encourage students to reflect on their experience and perceptions of GenAI and provide feedback to improve the course design. Because of this addition, we reduced each module to three weeks. This challenged us to combine weeks from the course maps in Phase 1. We also decided to limit each module to one discussion.

Since we moved the activities for week 2 to Module 2, we decided to concentrate Module 1 on prompt engineering skill-building and practice. In week 3, we planned for students to participate in a discussion by sharing prompts and outputs with their peers. We removed this discussion because we had a strong discussion in week 1, and week 4 contained an assignment about prompt engineering. We used ChatGPT to merge the concept for the discussion in week 3 with the assignment in week 4. The ChatGPT-generated text provided four options, one of which was the “Prompt Engineering Portfolio” (OpenAI, 2025). The Prompt Engineering Portfolio aligned the most with the original concept of the assignment we planned in week 4. We followed this output by prompting ChatGPT to create an assignment template. The ChatGPT-generated text provided an assignment template with instructions for students to describe their initial prompt creation, how the prompts were refined and improved, and an analysis and reflection. We analyzed the ChatGPT-generated template and realized the assignment lacked a hands-on, theory-to-practice connection and graduate-level rigor. Therefore, we prompted ChatGPT, “How could this assignment be extended?” The ChatGPT-generated text provided five options to “deepen the learning experience” (OpenAI, 2025). We decided to proceed with one of the five options ChatGPT-generated, the “Application Scenario Simulation,” which required the students to “select one of the refined prompts and create a short classroom activity, assignment, or lesson plan that incorporates it but changed the name to “Implementation Report” (OpenAI, 2025). This led to five versions of the assignment template after the design team prompted ChatGPT to provide word limit guidance, merge or alter sections, or change instructions. Consequently, these changes led to eight versions of the rubric from prompts aimed at appropriately allocating points to each criterion or changing the performance levels. Because of these changes to week 3, week 4 was left unchanged.

The process throughout Module 1 contained several decisions that impacted the entire design. For example, the readings selected for Module 1 informed the rest of the modules. After reviewing textbooks on artificial intelligence and education, the design team used literature from government education agencies such as the US Department of Education, UNESCO, and AI4K12 because they were relevant, accessible, and practical to our student population (AI4K12, 2020; UNESCO, 2023; U.S. Department of Education, 2023). Initially, we selected UNESCO's Guidance for Generative AI in Education and Research (2023). However, we decided it was too focused on policymaking and did not meet the needs of our students, who are currently in-service practitioners. Instead, we assigned John Spencer's (2022) A Beginner's Guide to AI in Education and the US Department of Education's (2023) Artificial Intelligence and the Future of Teaching and Learning. We decided that any assigned readings needed to be relevant for in-service teachers to use in their teaching practice. We also used ChatGPT to make this decision by prompting ChatGPT, “Make an argument for using this resource, A Beginner's Guide to Artificial Intelligence in Education by John Spencer versus the UNESCO Guidance for Generative AI in Education and Research.” The ChatGPT-generated text provided a comparison and suggested using both resources strategically throughout the course. However, we decided to use Spencer (2022) because it is practitioner-focused and supports the course learning objectives. Figure 2 displays the prompt and the initial portion of the output.

Figure 2

Prompt and Output with ChatGPT to Determine Module 1 Readings

In conclusion, Module 1 progressed from a four-week introduction to GenAI in education into a practical, skill-oriented three-week module focused on building GenAI literacy and preparing in-service teachers to use GenAI tools. The main design moves in Module 1 were determining the readings that would best serve our practitioner audience and removing some elements to align the module's focus with prompt engineering. Throughout this module, we learned that ChatGPT can provide numerous versions of activities, assignments, and rubrics based on learning objectives and course maps. We also discovered that we could enter passages from the readings into ChatGPT to ensure the activities and assignments support our view of alignment with the module and course learning outcomes. Figure 3 displays Module 1 as it appears in the learning management system (LMS).

Figure 3

Module 1 within the LMS

Module 2

Phase 1

In Phase 1, weeks 1 and 2 were designed to introduce using to support GenAI curriculum development. In week 1, students were expected to explore GenAI tools for personalized learning and differentiation. During week 2, we planned for students to learn how to use GenAI to streamline creating instructional materials. Week 3 focused on introducing ways to use GenAI to save time for communication, co-planning, and data analysis. Lastly, in week 4, we intended for students to create a comprehensive plan for using GenAI, selecting their preferred tools and strategies throughout the module. Throughout Phase 1, the design team struggled with differentiating between Module 2 – Teacher Use of GenAI and Module 3 – Student Use of GenAI. At times, the instructional materials, activities, and assignments overlapped. Therefore, we created a differentiating table to separate the two modules. This table is displayed in Table 1. In Module 2, students would learn how to use GenAI to support curriculum development, reduce administrative tasks, and support differentiation. In Module 3, students would learn strategies for teaching about GenAI and how to use GenAI with students across subject areas.

Table 1

Module Comparison Table for Module 2 and 3

Phase 2

In Phase 2 of the design process, we refined the module to span three weeks, aligning it with the structural changes implemented across all modules. In week 1, we added a reflective journal assignment that challenged students to critically examine a GenAI tool to use in their professional teaching practice to save time or differentiate learning. During week 2, we initially considered adding two discussions: one exploring using GenAI to teach state standards and another analyzing a scenario. We combined these into a single discussion that asked students to analyze how GenAI tools can enhance teacher efficiencies while supporting their instruction of state standards.

Week 3 culminated in an assignment titled “AI Integration and Efficiency Plan.” This was the most complex task in the module and required students to synthesize their learning from the previous two weeks. The assignment asked students to select 2–3GenAItools, develop a plan for streamlining tasks and personalizing instruction, and reflect on ethical considerations such as bias and student data privacy. Using ChatGPT, we generated the first version by prompting: “Design an assignment for graduate-level educators to create an GenAI integration plan that includes tool selection, task alignment, implementation timeline, and ethical review.” We then asked for a rubric with performance levels and detailed descriptors. We iteratively refined it to include five key areas: Personalizing Learning, Streamlining Tasks, Ethical Considerations, Clarity and Organization, and Flowchart or Diagram. We worked with ChatGPT to generate and refine four assignment drafts and six rubric versions to ensure alignment with learning objectives and rigor for graduate-level students.

Overall, Module 2 evolved significantly in Phase 2 through the iterative use of ChatGPT and feedback from the design team. The module benefited from the differentiation table we created in Phase 1, which separated teacher versus student use of GenAI. ChatGPT offered drafts and creative variations throughout the design process that we could refine. By the end of the module, students develop plans to integrate GenAI into their practice, emphasizing efficiency and differentiation and laying the foundation for deeper student-centered applications in Module 3.

Module 3

Phase 1

In Phase 1, our team structured Module 3 - Student Use of GenAI as a four-week module focused on using GenAI with students across different subject areas. In week 1, the course map outlined that learners would complete readings on GenAI ethics and write a discussion or reflection expressing their concerns about GenAI use with students. Weeks 2 and 3 were structured to allow students to explore GenAI's open educational resources and lesson plan repositories. We planned for students to share their findings in a discussion and, in week 4, complete an assignment that pairs GenAI tools with specific teaching strategies.

Phase 2

In Phase 2, we first had to decide how to consolidate the module into three weeks. Weeks 2 and 3 in the course map were conceptualized as “Having students use GenAI in the Classroom, Part 1 and 2.” Considering the addition of Module 5, we decided to combine these weeks. Additionally, because we decided to limit each module to one discussion, we had to reevaluate the discussions in week 1 and week 2. The discussion we planned in week 1 was similar to Module 1, and the option in week 2 was limited in scope. As we deliberated, we realized we wanted the discussion to focus on the ethical considerations surrounding student use of AI. Therefore, we prompted ChatGPT to create a discussion focused on students' ethical use of GenAI. The ChatGPT-generated text provided a case study that challenged students to propose ethical guidelines for student use of GenAI (OpenAI, 2025). Apart from changing the language, we adopted the concept of this discussion for Module 3 because it aligned with what we wanted students to accomplish in this module.

Next, we turned to designing and combining week 2 and week 3 and refining Assignment 3. In Phase 1, we intended for students to explore GenAI open educational resources, identify three tools and strategies, and reflect on how they would incorporate them into their teaching practices. We decided that this assignment could be improved and considered using the AI4K12 Guidelines for teaching AI in K-12 classrooms as the foundation for the assignment (AI4K12, 2020). The AI4K12 Guidelines focus on AI's “5 Big Ideas”: perception, representation and reasoning, learning, natural interaction, and societal impact. We believe these ideas would benefit our students as they grow and develop into GenAI leaders in their context. Therefore, we prompted ChatGPT, “What if for Assignment 3, I have students analyze the AI4K12 Big Ideas and then use an GenAI tool to create a lesson for their teaching context based on one of the “Big Ideas.” The ChatGPT-generated text provided a template, and we refined several iterations of the assignment before abandoning the idea because we realized that the AI4K12 Big Ideas would serve the module better as a reading for students to engage with before the assignment. More importantly, we wanted students to find ways to adopt and embed GenAI rapidly into their current teaching practices, not necessarily teach students about the mechanics of GenAI. This decision was again a return to our initial course learning outcome of creating a course focused on practical classroom implementation of GenAI. We prompted ChatGPT to create a version of the assignment that challenges students to design a lesson plan using their existing curriculum or instructional materials as a starting point, where students would interact directly with GenAI tools.

Throughout the entire design process, Assignment 3 required the most prompting and refinement with ChatGPT. Many of these changes were because we abandoned using AI4K12 in the assignment and the ChatGPT-generated outputs did not align with the course objectives. Overall, the template and structure of the assignment went through seventeen iterations, and we refined our prompt a total of forty-seven times to request various changes and tweaks to each iteration because each version of the assignment ChatGPT created did not align with the needs of our learner population. We prompted ChatGPT to make minor adjustments, such as changing the name of the assignment, and significant modifications, such as removing entire components or creating blended versions of the various iterations. We also prompted ChatGPT to analyze the current iteration to ensure alignment with the course and module learning outcomes, critique each variation, and argue which was the strongest based on the module learning outcomes. As we refined each prompt with ChatGPT, we continually tried to return the assignment toward a practical, hands-on activity relevant to our students. The prompt that finally led us to abandon AI4K12 and arrive closer to the final iteration was, “Many of these educators are not necessarily teaching students about AI. However, I want this assignment to support educators in using GenAI for what they already do in the classroom. How could this assignment reflect that?” This prompt started the round of iterations that led to creating an assignment focused on practically integrating GenAI using existing curriculum or instructional materials. An image of this prompt and the beginning of the output is shown in Figure 4.

Figure 4

Assignment 3 Prompt and Output with ChatGPT

Overall, the main design moves in Module 3 were condensing the discussions into a single topic focused on analyzing a case study and proposing guidelines for ethical student use of GenAI in the classroom. Additionally, Assignment 3 shifted from using one of the Big Ideas from AI4K12 into a practical lesson plan built on embedding GenAI into existing instructional materials or curriculum. Throughout this module, we realized we could prompt ChatGPT to interrogate and critique what it creates. This is critical when designing instructional materials using ChatGPT to ensure the highest quality resources and alignment with learning objectives. In conclusion, Module 3 grew from a four-week exploration of open educational resources in Phase 1 into a targeted three-week module centered on embedding GenAI using existing curriculum and creating classroom guidelines for the ethical use of GenAI in the classroom in Phase 2. Figure 5 displays Module 3 as it appears in the LMS.

Figure 5

Module 3 within the LMS

Module 4

Phase 1

In Phase 1, we planned for Module 4 to prepare educators to become GenAI leaders within their educational communities. During week 1, students would examine the ethical considerations and implications of using GenAI in teaching and learning and create policies for using GenAI in weeks 2 and 3. In week 2, they would design a policy for their classroom and their entire school in week 3. Lastly, we planned for students to develop professional learning materials in week 4.

Phase 2

In Phase 2, we first consolidated the four weeks into three by combining weeks 2 and 3, which were devoted to creating policies. Initially, we planned separate activities for drafting policies and GenAI tool curation. However, after considering the practical and actionable activities we developed in previous modules, we combined them into a single discussion. To support this discussion, we used ChatGPT to generate a table that students could use to evaluate and share GenAI tools. We prompted ChatGPT, “Design a table that allows educators to compare educational GenAI tools by input, process, model type, output, and classroom use.” After modifying the output, we included it as a required format for the discussion to support a more systematic evaluation of tools.

We also added a reflective journal assignment that allowed students to analyze the ethical and practical implications of using centralized, customized, and hybrid GenAI models in education. However, we noticed that we had included limited information on the types of GenAI models in the readings. Therefore, we added two additional resources to week 1 to better support students completing the assignment. One resource was developed by the design team that defined each model and provided examples. The other was an article from a reputable GenAI industry publication. Another modification to Phase 2 was the addition of the TeachAI Toolkit and selected readings from the US Department of Education and ISTE, which we chose for their focus on GenAI policies and accessibility for teachers with diverse backgrounds (ISTE, 2025; TeachAI, 2025; U.S. Department of Education, 2023).

Lastly, in Phase 1, we debated whether students should develop a school-wide GenAI policy or create a professional development plan for the culminating assignment. We decided that students should create a professional development plan because most students would have more control and opportunities to deliver professional development than creating official school policies. Furthermore, the professional development plan assignment would require students to apply pedagogical and policy knowledge gained throughout the course. This aligned with the course learning objective to emphasize leadership, not just implementation.

In designing Module 4, we learned that ChatGPT can assist with scaffolding our instruction by generating instructional materials and exemplars. However, the design team needed to evaluate each resource to ensure it met the rigor expected in graduate education and the needs of our practitioner audience. Designing with ChatGPT in this module required continuous iteration, oversight, and analysis.

Module 5

The course's fifth and final module challenges learners to reflect on their journey with GenAI in education. As mentioned in the description for Module 1, Module 5 was designed entirely in Phase 2. Initially, we planned for a simple final reflection in a discussion. However, as we deliberated, the final assignment grew into a two-part assignment. First, learners would consider how their perceptions of GenAI have shifted throughout the course and how the insights they have gained shape their future teaching practice, and second, offer recommendations for course enhancements and provide additional feedback. Table 2 provides an overview of each module including how the designed team used ChatGPT and how the learners will use GenAI in each module.

Table 2

Course Overview

Evaluation & Implementation

We utilized a formative evaluation approach, specifically looking at the product (Rothwell & Kazanas, 2008). We utilized two different expert perspectives to review the course. First, we had an instructional designer based within the College of Education review this from a content and delivery perspective. In addition, we had a former student of our M.Ed. program, a current P12 practitioner, review the course from a delivery and usability perspective. Their initial review identified areas of concern regarding navigation or implicit instructions.

Through these two expert reviews, three key points were identified: two points were actionable feedback, while the third was more of an acknowledgment of an ongoing challenge of the course. The first element related to the accessibility and consistency across modules within the course. From an accessibility standpoint, the feedback highlighted features of the existing course and included suggestions that would help the course meet upcoming changes to accessibility guidelines. This included things like specific fonts and limiting particular emphases. In terms of consistency, this level of feedback identified minor mistakes made in the development process (e.g., a broken link or a missing checklist). The second element focused on suggestions related to peer-to-peer engagement, such as providing opportunities for the learners to co-create resources that would be beneficial within each module. The final element raised was the challenge of keeping the course updated, as it is situated within an emerging field, whereby access and use of these tools change rapidly. This challenge was especially noted in the discussions regarding ethics and policy, as there is a possibility that many districts may implement restrictive policies at some point in the future.

The course was taught for the first time in Spring 2025 across the 16-week semester and is planned on being offered each semester following; however, it will also be offered in an approximately eight-week version during the summer semester. The course is designed and coded as 100% asynchronous; however, there are ample opportunities for synchronous interaction, including virtual meetings.

Reflections on the Design Process

Design Philosophy

This reflection section will highlight the experience of designing the course and using ChatGPT throughout the design process. Throughout the course design, we decided to refrain from suggesting one GenAI tool or chatbot over another. We made this decision for several reasons. Firstly, throughout the course, we continually focus on strategies over tools. We want to empower students to learn how to use and navigate all GenAI tools by giving them a general understanding of their mechanics and functionality. Second, because of the rapidly evolving nature of GenAI, we wanted to ensure that the course was not already out of date by the end of the semester. If we focused on strategy over tools, this would provide a certain level of stability from semester to semester. Third, many GenAI tools now require a paid subscription to utilize advanced models capable of higher-level reasoning and logic. Students in the course can access Microsoft Copilot through their university Microsoft account.

Additionally, some students may have access to GenAI tools provided by their district, such as MagicSchoolAI or others they have purchased. We informed students that they had access to Copilot but were not required to use it. Nevertheless, we wanted to focus on strategies over tools.

Even before the formative evaluation feedback, the designers anticipated that this course would be challenging to keep up to date because of GenAI's rapid growth and development. The designers focused less on the technical aspects of using GenAI. Instead, they focused on equipping learners with the specific strategies, skills, and knowledge to use GenAI and support its adoption in their teaching context. Equity and accessibility-related challenges are also evident, considering that many advanced features, including varying intelligence and reasoning, are increasingly limited to paid monthly subscribers. Although this course does not require a subscription to use many popular GenAI tools, learners may be financially limited to using more robust GenAI tools in the future.

Technical Constraints

The design of this course was not without some technical limitations that prevented the designers from taking specific actions, such as being integrated into learning platforms to provide personalization for learners, which would change the types of activities or strategies that students are exposed to (Heeg & Avraamidou, 2023; Yue et al., 2021). Given the limitations of our LMS, this was something that could not currently be explored. In future course iterations, the designers would like to explore how to overcome this limitation and integrate tools on a much deeper level to support the learner's progression through the course.

ChatGPT in Course Design

ChatGPT supported the entire instructional design process, from supporting the development of course maps and learning objectives to generating instructional materials, assignments, and rubrics. Using ChatGPT allowed for rapid redesign and iteration to ensure relevancy and authenticity for learners and unique possibilities to transform assessment practices by rendering alternative assignments that require students to apply their knowledge to a specific context or problem. Following the pattern of emerging technologies and previous innovations in instructional technology, ChatGPT presents unique affordances, constraints, and difficulties that were apparent during this design case. In terms of affordances, ChatGPT never grew tired or weary of creating or redesigning instructional materials. As discussed in the development of each module, most of the materials developed for this course were created using ChatGPT and refined through subsequent prompts. Before the final draft, some assignments, discussions, and assessments had fifteen to twenty iterations containing minor and significant changes. The course designers reviewed, refined, and iterated on all ChatGPT-generated text.

Furthermore, the use of ChatGPT followed the emerging capabilities and refinement of ChatGPT models. In Phase 1, some design team members used the free version of ChatGPT-4 while others used the ChatGPT Plus access. In Phase 2, the design team utilized ChatGPT Plus with access to advanced models for reasoning and logic. At times, our work with ChatGPT was not always linear. As various iterations of assignments and instructional materials were created, they were entered back into ChatGPT to ensure alignment with the course and module learning objectives. Our prompts at the beginning of Phases 1 and 2 followed “system” and “user” roles and general guidance from OpenAI about prompt engineering (OpenAI, 2025). For example, the following system message provided a context for how the ChatGPT should behave and respond: “Act as if you are an expert instructional design who is highly skilled in instructional design and technology for graduate-level online courses.” However, as the design process progressed, we realized the outputs from ChatGPT were improving because of the data and materials we entered in prompts and the “memory” feature provided in an update in the initial design phases and improved throughout the entire design process. Throughout this design case, we showcased the types of prompts used and the outputs we received to design instructional materials, assignments, and rubrics.

The quality and depth of ChatGPT-generated outputs varied (e.g., Assignment 3). This made us curious about the “jagged technological frontier” of GenAI tools within instructional design (Dell’Acqua et al., 2023). ChatGPT handled some prompts with expert precision and detail throughout the design process. However, the quality and characteristics of the outputs could vary wildly from prompts requiring similar knowledge and skill. This made us also curious about the average output from large language models (LLM) when prompted to complete an instructional design task. LLM models are designed to provide outputs that align with the training data's most common patterns and trends, essentially an average representation of what it has learned. Therefore, if we wanted to develop unique and creative learning activities, assignments, or assessments, we had to create prompts that encouraged ChatGPT to do so. If we relied on ChatGPT without evaluating, refining, and iterating the outputs, we faced the likelihood of ChatGPT stifling innovation and creativity.

ChatGPT also does not truly understand how to teach in-service educators to use GenAI in their teaching practices. Therefore, despite the high writing quality and convincing authority in the outputs, an expert needed to analyze and critique each piece of generated material manually. This reflects our desire to maintain a “human in the loop” (US Department of Education, 2023). When reflecting on the difficulties of using ChatGPT, we often wondered if students would be offended that course materials were developed using a GenAI chatbot. From our perspective, ChatGPT has access to or has learned about instructional design, education, and itself as a GenAI tool through its vast training data. Therefore, our design team thought it would be a mistake not to leverage this new and novel synthetic expert to ensure our course is of the highest quality (Ringel, 2023). The medical field is experiencing a parallel situation in diagnosing and determining patient care. New LLM models such as ChatGPT-o1 can now diagnose patients with surprising accuracy, and many physicians are using them to support their clinical decision-making (Goh et al., 2024; Kanjee et al., 2023). Considering these uses, we thought it was essential to use ChatGPT to generate second opinions and ensure learners receive the highest quality instruction possible.

Conclusion

As learners participate in the course, we anticipate directly impacting their professional practice. The course intends for learners to become change agents within their organizations who are knowledgeable of GenAI tools and prepared to have challenging conversations around its implementation at the classroom level (i.e., with their learners) and the organizational level (i.e., with their peers and school leaders).

This design case outlined the development of a graduate course designed to equip educators with the skills and dispositions to effectively integrate and deploy GenAI into teaching and learning practices, enhancing the educational experience for preschool through 12th-grade students. This course aims to address a critical need in higher education by designing educator-focused training to maximize the potential of GenAI within P12 settings. At the same time, it highlights the shortcomings and concerns of using these tools.

ChatGPT was used throughout the course development to create and refine drafts of instructional materials, including discussion prompts, assignment templates, and rubrics. Throughout this process, the course designers identified the affordances and constraints of using ChatGPT for practical, applied instructional design. ChatGPT frequently produced high-quality instructional materials aligned with the course and module learning objectives. However, the designers often redirected and corrected ChatGPT to ensure the materials aligned with the course’s objective and vision. Additionally, the design team finalized and iterated all instructional materials created from ChatGPT-generated text. From this design case experience, we continue to approach ChatGPT with cautious optimism. As we continue incorporating ChatGPT into instructional design, we will evaluate and critique all outputs to ensure alignment with course and module learning objectives and maintain high-quality instruction for our learners.

References

AI4K12. (2020). AI4K12. https://ai4k12.org

Boling, E. (2010). The need for design cases: Disseminating design knowledge. International Journal of Designs for Learning, 1, 1–1. http://scholarworks.iu.edu/journals/index.php/ijdl/index

Dell’Acqua, F., Mcfowland, E., Mollick, E., Lifshitz-Assaf, H., Kellogg, K. C., Rajendran, S., Krayer, L., Candelon, F., & Lakhani, K. R. (2023). Navigating the jagged technological frontier: Field experimental evidence of the effects of AI on knowledge worker productivity and quality.

Howard, C. D. (2011). Writing and rewriting the instructional design case: A view from two sides. International Journal of Designs for Learning, 2(1), 40–55. http://scholarworks.iu.edu/journals/index.php/ijdl/index

ISTE. (2025). International Society for Technology in Education. https://iste.org/

Lee, S. J., & Kwon, K. (2024). A systematic review of AI education in K-12 classrooms from 2018 to 2023: Topics, strategies, and learning outcomes. Computers and Education: Artificial Intelligence, 6, 100211. https://doi.org/10.1016/j.caeai.2024.100211

Ringel, D. M. (2023). Creating synthetic experts with generative artificial intelligence. Kenan Institute of Private Enterprise Research Paper.

Rothwell, W. J., & Kazanas, H. C. (2008). Mastering the instructional design process: A systematic approach (4th ed.). Pfeiffer.

Seufert, S., Guggemos, J., & Sailer, M. (2021). Technology-related knowledge, skills, and attitudes of pre- and in-service teachers: The current situation and emerging trends. Computers in Human Behavior, 115. https://doi.org/10.1016/j.chb.2020.106552

Smith, K. M. (2010). Producing the rigorous design case. International Journal of Designs for Learning, 1(1), 9–20. http://scholarworks.iu.edu/journals/index.php/ijdl/index.

Tate, T. P., Doroudi, S., Ritchie, D., Xu, Y., & Warschauer, M. (2023). Educational research and AI-generated writing: Confronting the coming tsunami. EdArXiv. https://doi.org/10.35542/osf.io/4mec3

TeachAI. (2025). AI Guidance for Schools Toolkit. teachai.org/toolkit

Trust, T., Whalen, J., & Mouza, C. (2023). Editorial: ChatGPT: Challenges, opportunities, and implications for teacher education. Contemporary Issues in Technology and Teacher Education, 23(1), 1–23.

UNESCO. (2023). Guidance for generative AI in education and research. https://doi.org/10.54675/EWZM9535

U.S. Department of Education. (2023). Artificial intelligence and the future of teaching and learning: Insights and recommendations.

U.S. Department of Education. (2024). National Educational Technology Plan.

Appendix

AI Needs Assessment for IT Program

How interested are you in learning about AI?

Not interested 1 – 2 – 3 – 4 – 5 Very interested

How knowledgeable are you about AI?

I am a complete novice 1 – 2 – 3 – 4 – 5 Extremely knowledgeable

How would you rate your skills with AI?

I’m an AI Beginner 1 – 2 – 3 – 4 – 5I'mm an AI Expert

Have you personally used any of the following generative artificial intelligence (AI) tools?

ChatGPT

Gemini

Scikit

Learn

Dalle-2

Midjourney

Crayion

Copilot

Other

What is your general perception of Artificial Intelligence (AI) technology related to your profession?

Extremely negative 1 – 2 – 3 – 4 – 5 Extremely positive

What about AI is most important to you? Check all that apply.

Policy surrounding AI

How I can use AI to make my job easier as a teacher

How I can get my students to use AI responsibly

Using AI in my non-classroom position

Other:

Have you received professional development from your district related to AI and teaching/learning?

No

Yes

Other:

If any, what are your concerns about implementing AI into your classroom or position?

What would you need to implement AI into your classroom or position?

How familiar are you with free tools that may be available to you that would allow you to utilize AI technologies?

What about AI and/or AI tools would you like to learn about?

If you were to use AI in your teaching, how might you imagine using it?

Is there anything else you would like to add about this topic?