On this page

Synthesis: Three University of Michigan designers built a three-course short series on Coursera in four months, aimed at learning design professionals rather than at students. The series treats generative AI as a design partner across the design workflow: objectives, outlines, assessments, evaluation plans. Three problems shaped every choice: practitioners needed research translated into practice, the tools would change under the course, and the topic attracts polarized claims. The authors answer with translational framing, a split between ephemeral and evergreen content, and priming activities for contested questions. The case reports no learner outcomes; the authors say they still do not know how the material will age.

Key Findings

  1. The audience was practitioners, and the content had to translate. The designers cast themselves in Norman's role of translational developers, converting research on AI and learning design into demonstrations a working designer could apply the same week.
  2. Relevance was designed in, not assumed. Objectives were written for enduring understanding, and content was sorted into ephemeral material (specific tools, interfaces, prompts) and evergreen material (design principles, evaluation questions). The split decides what gets updated and what survives a model release.
  3. Polarized claims were handled with priming, not avoidance. Rather than settle the ethics and labor arguments, the series gives learners structured activities and priming to articulate a position and compare it with peers, so the disagreement becomes part of the curriculum.
  4. The format constrained the design. Coursera short courses run one to five hours, which pushed the series toward consolidated demonstrations and away from extended projects; the designers treat that as a design parameter, not a limitation to lament.
  5. The design case is also a reflection on design practice. The authors report that the three problems generalize beyond GenAI: translating research, maintaining relevance under rapid change, and teaching contested topics are recurring conditions of learning design work.

What a "design partner" course has to decide

The phrase covers a wide range of practice, and the series narrows it by walking the design workflow: writing objectives, drafting outlines, generating assessment items, planning evaluation. At each step the question is what the designer keeps and what the model produces, which is the judgment the course is trying to teach. The ephemeral/evergreen distinction is the mechanism that keeps this honest over time: a demonstration of a specific tool ages in months, while a question about how to check whether a generated assessment measures the objective does not.

What this means for practice

  • Instructional designers. Sort your AI material into tool-specific and principle-level content before you write it. The sort tells you what needs a maintenance cycle and what does not.
  • Instructional designers. Treat contested claims as an activity rather than a disclaimer. Learners who have to state and defend a position engage with the disagreement; a paragraph warning about hype does not.
  • Faculty developers. The translational stance is a job description. Someone has to read the research and turn it into a demonstration, and a course that skips that step becomes a tool tour.
  • Researchers. The series has run for over a year with thousands of enrollments, and the authors name completion, assessment and forum data as the evaluation they have not yet done. That is an open opportunity, not a result.

Limitations

  • No outcome data: no completion analysis, no assessment results, no comparison with another course design. The authors describe learner feedback as positive and the series as highly rated without reporting it.
  • The account is written by the designers and instructors of the series, which makes it a design case rather than an appraisal.
  • The course targets learning design professionals in one MOOC platform, with a short-course format that limits project work; the design choices may not transfer to a semester course.
  • The authors state that they do not yet know how the material will age, which is the central risk of a course about a fast-moving technology.

Citation

Quintana, R. M., Quintana, C., & McCurry, M. (2025). Unpacking the design of an online course series: Generative AI as a learning design partner. International Journal of Designs for Learning, 16(2), 159–169. https://doi.org/10.14434/ijdl.v16i2.42086

Embed this page

Copy the code below to embed this page in a learning management system or other website. The embedded version shows only the page content — it hides the site header, navigation, and footer.