Assistant Professor Tufts University Medford, Massachusetts, United States
Introduction: First-year engineering design classes are crucial for engaging students and fostering a commitment to a biomedical engineering path early in undergraduate education. To enhance inclusivity and accessibility in these foundational courses, Universal Design for Learning (UDL) can be an effective framework to accommodate diverse student needs. However, integrating UDL strategies can prove challenging due to logistical burdens including resource constraints, time consumption, and perceived complexity. Generative AI offers significant advantages for creating diverse and personalized teaching content in ways that align closely with UDL principles while reducing instructor fatigue in assessing current course content and implementing changes. Here I present the development of a first year course, Sci-Fi Bioengineering, and UDL-driven course modifications driven by generative AI to enhance design content, diversify assessments, and increase peer-to-peer engagement.
Materials and
Methods: “Sci-Fi Bioengineering” is a semester long, lab-free course that uses 5 classic and contemporary sci-fi films as a framework to discuss advanced biomedical technologies and their societal and bioethical contexts. The course consists of several key components: film screenings to identify engineering design parameters, lectures and full-class discussions on current science and ethical issues, journal club discussions of research articles, group presentations on artificial organ technologies, rapid mini debates, and final podcast projects where students analyze sci-fi products in terms of engineering design, current research, and ethical considerations. ChatGPT (GPT4) was used to analyze formal course evaluations from the first iteration of the course to identify strengths and areas for improvement. Instructional “pinch-points” were identified around curriculum, assessment, instruction, and logistics to develop “+1” UDL strategies (Fig 1A) for the second, larger (30 vs. 20 students) iteration of the course. ChatGPT (GPT-4), MidJourney, and BingChat (Microsoft Co-Pilot) were used for content creation and incorporated into in-class projects. These included diversified film assignments and online discussion boards aligned with student interests identified in a pre-course survey, allowing students to choose their topics and broadening course material. Additionally, "design fiction" activities were developed where students iteratively created products and business pitches based on a sci-fi scenario at the beginning and end of the course to integrate learned materials with a “hands-on” component. Anonymous post-course surveys were used to assess learning outcomes in both iterations, and in the second to solicit targeted feedback on the success of novel UDL components.
Results, Conclusions, and Discussions: Mid- and post-course surveys in both iterations of Sci-Fi Bioengineering show the vast majority of students (>85%) self-report “slight” to “very improved” understanding of all course content, including principles of engineering design, engineering vs. basic science, genetics & gene editing, immunology, peer review and types of scientific communications, artificial organ technologies, laboratory research regulations and intellectual property, stem cell biology, cloning and regenerative medicine, neurophysiology and neurotechnology, and contemporary biomedical issues. Compared to the first iteration, a larger proportion of students report “very improved” understanding of “principles of engineering design” (Fig. 1B) and “engineering vs. basic science”, demonstrating the benefit of (+1) UDL-based instructional changes on engineering content. Specific survey questions indicate that (+1) assignments and (+1) online discussion boards positively contributed to student engagement with course material in a variety of ways (Fig 1C), and that design fiction activities specifically improved perceived student learning around engineering design (Fig. 1D).Qualitative feedback demonstrated that UDL strategies were not noticeably added to the course but stood out as impactful components. Open-ended prompts also show that UDL strategies contributed to student enthusiasm and appreciation for course structure and balance. From an instructor's perspective, UDL eased logistics by reducing emails regarding absences or confusion due to recorded lectures and online forums, enhancing student ownership through self-grading for participation, organically broadening course materials on discussion boards, decreasing stress-related office visits, and fostering greater student-to-student engagement despite fewer faculty-student connections in a larger class setting. Although design courses vary widely in content, student demographics, and resources, we anticipate that some of these curricular elements and instructional strategies can be effectively applied and scaled to incorporate UDL, even in small ways, to enhance the inclusivity and diversity of our student bodies, while simultaneously nurturing student interests and demonstrating the practical relevance of biomedical engineering to their lives.
