Barriers that Programming Instructors Face While Performing Emergency Pedagogical Design to Shape Student-AI Interactions with Generative AI Tools

Paper Title

Barriers that Programming Instructors Face While Performing Emergency Pedagogical Design to Shape Student-AI Interactions with Generative AI Tools

Publication Info

• Topic area: Challenges and adaptations in computing education due to generative AI tools.
• Keywords: Generative AI, computing education, emergency pedagogical design, student-AI interaction, HCI, course redesign, assessment, policy, resources.

Background and Problem

• Problem / challenge: Instructors face challenges adapting courses to widespread student use of generative AI tools, with limited guidance, resources, and control over commercial AI interfaces.
• Significance: Generative AI tools are pervasive and impact student learning, necessitating course redesigns to ensure productive student-AI interactions while addressing equity and effectiveness concerns.
• Motivation and related work: Prior research has explored GenAI’s capabilities and instructor perceptions but lacks insights into practical course adaptations. This paper builds on this gap by studying instructors’ real-world efforts and barriers.

Solution

• Proposed approach: Emergency pedagogical design—reactive, indirect efforts by instructors to shape student-AI interactions under time pressure and partial observability.
• Novelty:
  1. Conceptualizing emergency pedagogical design as a distinct design setting for HCI.
  2. Documenting five barriers instructors face: fragmented buy-in, policy crosswinds, implementation challenges, assessment misfit, and lack of resources.
  3. Offering recommendations for HCI researchers, institutions, and funders to support instructors in adapting courses to GenAI.
• Procedure and key techniques:
  • Conducted interviews (n=13) with computing instructors who had updated course materials for GenAI.
  • Surveyed computing faculty (n=169) to capture broader perspectives.
  • Used inductive thematic analysis to identify barriers and patterns in emergency pedagogical design.

Results

• Concrete findings:
  • 81% of surveyed instructors were open to adopting GenAI, but only 28% perceived departmental support.
  • 85% were likely to change curricula due to GenAI, yet only 26% felt institutional policies positively impacted their efforts.
  • 53% lacked sufficient resources, with MSI instructors reporting heavier teaching loads and fewer resources.
• Advantage over baselines: Highlights practical barriers and actionable insights that prior work on instructor perceptions and GenAI capabilities did not address.
• Experiments / evaluation:
  • Interviews focused on instructors who had implemented changes beyond policy edits.
  • Survey captured diverse faculty perspectives, including those at MSIs and HBCUs.
  • Analysis emphasized barriers and instructor needs for effective course redesigns.
• Limitations and future work:
  • Findings may not generalize beyond computing education or U.S.-based contexts.
  • Self-reported data is subject to recall and social-desirability biases.
  • Future work should explore effective student-AI interactions, equity concerns, and scalable tools for instructors.

Summary

This paper introduces the concept of emergency pedagogical design to describe instructors’ reactive efforts to adapt courses for productive student-AI interactions amidst widespread generative AI use. Through interviews and surveys, the authors identified five barriers—fragmented buy-in, policy crosswinds, implementation challenges, assessment misfit, and lack of resources—that constrain these efforts. The study emphasizes the need for HCI research, institutional support, and funding to address these barriers and enable scalable, equitable course redesigns. Findings provide actionable insights for improving student-AI interactions and supporting instructors in navigating the challenges of integrating GenAI into education.