The Penn AI Pedagogy Initiative (PAPI) addresses a key institutional challenge: while universities increasingly invest in faculty development and digital capacity-building, the adoption of innovative pedagogies remains limited to early adopters. To enable responsible, meaningful, and scalable AI integration in education, our proposal establishes a structured methodology that directly involves faculty and students in co-designing, testing, and evaluating AI-enhanced pedagogical practices.
At the heart of our project lies a key principle: AI in education must be co-constructed, not imposed. This philosophy is operationalized through an iterative, research-informed process for sustainable and agile innovations. Faculty and students will work collaboratively to prototype AI-enhanced learning materials and activities that reflect authentic classroom practices. This ensures that AI is not adopted for its novelty but is thoughtfully applied to deepen learning, foster engagement, and amplify the uniquely human aspects of teaching. We intend to keep pace with the rapid developments in AI and adopt/develop the right pedagogical tools for different fields with our multi-disciplinary team.
The Institutional Imperative
The rapid proliferation of Generative AI has placed global universities under significant pressure to modernize instructional frameworks. However, research indicates that traditional capacity-building efforts—such as isolated workshops or voluntary faculty courses—often fail to drive institution-wide behavioral change, typically reaching only "early adopters". As noted by Hughes et al. (2025), reimagining higher education requires navigating complex adoption challenges that go beyond technical access. Furthermore, Kottmann et al. (2024) highlight that the innovation behavior of instructors is frequently hindered by the speed of technological evolution, leading to a "retrofitting trap" where advanced tools are forced into legacy instructional models. The Penn AI Pedagogy Initiative addresses this gap by providing a low-risk, high-support environment for faculty to move from AI anxiety to pedagogical agency.
Theoretical Framework:
Our approach is grounded in the principle that responsible AI integration must be context-specific and grounded in the unique "verbs" and “grammar” of each academic discipline. Following the recommendations of Zawacki-Richter et al. (2019), we emphasize that AI applications in higher education must be purposefully integrated to enhance, rather than replace, human-centric teaching.
The "Synapse" Methodology Central to our framework is the Synapse Model, which facilitates a high-fidelity collaboration between three distinct roles:
- The Disciplinary Explorer (Faculty): Provides domain-specific expertise and identifies "Stuck Points"—complex concepts where student cognitive load often peaks.
- The Cognitive Architect (Penn GSE Research Team): Brings expertise in learning sciences, instructional design, and pedagogical scaffolding.
- The Technical Catalyst (Interdisciplinary Student Partner): Offers domain-relevant technical fluency from schools such as Penn Engineering, Wharton, or Penn Arts and Sciences. This triad ensures that AI interventions are informed by both the Learning Sciences and Subject Matter Expertise.
Operational Strategy:
The Four-Stage Iterative Loop The initiative utilizes a structured 16-week cycle designed to move from theoretical inquiry to classroom-ready prototypes:
Stage 1: Classroom Observation (6 Weeks)
A deep-dive phase involving literature reviews, feedback analysis, and the identification of "Stuck Points".
Stage 2: Agile Co-Design (4 Weeks)
Collaborative prototyping of Minimum Viable Products (MVPs), such as digital personas for dialogic inquiry or rapid-automated-feedback systems.
Stage 3: Iterative Testing (4 Weeks)
Implementation of prototypes in live classroom environments to evaluate "intellectual throughput" and student engagement.
Stage 4: Evaluation and Dissemination (2 Weeks)
Summative assessment and the contribution of teaching notes and prototypes to the Penn AI Pedagogy Repository.
Strategic Objectives and Global Reach
The ultimate output is the creation of a living digital library—the Penn AI Pedagogy Repository—which documents proven strategies and adaptable models for K-16 institutions worldwide.
How to Get Involved
We welcome faculty from all universities and schools, at any level of prior experience—from those new to AI to regular users—to participate by integrating a course into this design-based, co-research model. You do not need to be an AI expert; the research team provides structured support throughout the design cycle.
We understand that some of the most innovative pedagogical work happens in quiet corners of the university. If you have a colleague who is currently experimenting with AI in their classroom—or someone whose teaching philosophy would suit this kind of collaborative, design-based exploration—we want to hear about them. Please feel free to fill out the same interest form to nominate a fellow faculty member. We will reach out to them to discuss and explore their interest.
What We’re Looking For
Pedagogically Grounded Exploration: Faculty interested in exploring AI integration in thoughtful, grounded ways.
- Active Courses: Instructors teaching courses during the upcoming Fall Semester 2026
- Collaborative Spirit: Openness to collaborative co-design with trained student research assistants.
- Commitment to Reflective Practice: Willingness to participate in systematic documentation and share insights with the broader community.
Initial Planning (60–90 min): A foundational meeting to discuss your current teaching philosophy, specific learning objectives, and the "pain points" in student engagement where AI might offer a pedagogical solution.
Classroom Observations (3 sessions): Allowing the research team to conduct non-disruptive observations aligned with your schedule to understand your instructional style.
Material Access: Providing relevant items such as syllabi, assignment prompts, rubrics, and selected Canvas pages.
Co-Design Conversations (2 sessions): 45–60 minute meetings to shape and refine AI-enhanced activities and materials (e.g., custom tutors, digital personas, or simulations).
Prototype Review: Providing disciplinary feedback on the low-fidelity prototypes created by the student team to ensure they align with your course standards.
Classroom Implementation: Facilitating the co-designed pedagogical activity during your scheduled class time.
Brief Reflections: Completing short check-ins or surveys at key points to document evolving thinking and gather feedback on student interactions.
Final Debrief: A brief closing conversation to evaluate the intervention’s impact on learning and engagement.
Repository Review: Reviewing and lightly editing a short course-level summary for the Penn AI Pedagogy Repository.
Faculty Autonomy & Ethics
- Instructional Agency: Faculty retain full autonomy and may adjust the activities as they see fit to their instructional goals
- Professional Partnership: The research team acts as respectful partners, aiming to minimize disruption and provide practical design support.
- Confidentiality: All reporting and dissemination use pseudonyms or de-identified data unless explicit permission is granted.
Get Involved
Are you a faculty member that would like to get involved, learn more about how to get involved or complete our interest form.
Contact Us
For questions about PAPI, please contact Nam Nguyen (he/him).
-
Associate Professor, Faculty Director of Learning Analytics and Artificial Intelligence Program, Penn GSE
-
Nam Nguyen
Senior Research Manager, AI & Education, Penn GSE