← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

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

Sam Lau , Kianoosh Boroojeni , Harry Keeling , Jenn Marroquin

DOI PDF Program page Page JSON

This looks like a strong CHI framing paper because it names a real instructional condition that many educators are already living through but that HCI has not clearly conceptualized. The mixed-method scope gives the framing credibility, and the five-barrier structure makes the contribution usable. Its main weakness in this review packet is not the idea itself, but the lack of access to body sections needed to inspect methodological rigor and explicit author-stated limitations.

Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Review Summary

The paper’s strongest contribution is that it converts a diffuse, reactive educational response to generative AI into a recognizable HCI problem setting. Instead of treating instructors as straightforward adopters of AI tools or as policy enforcers, it argues that they are engaged in “emergency pedagogical design”: reactive, indirect efforts to shape student-AI interactions while lacking control over the commercial interfaces students actually use. That is a meaningful conceptual move because it shifts the locus of design from interface construction to constrained pedagogical adaptation. In other words, the paper asks HCI to take seriously the labor of redesigning courses, assessments, and classroom norms under conditions of uncertainty, partial observability, and institutional ambiguity. The abstract also suggests that this is not merely a rhetorical reframing. The authors ground the concept in interviews with 13 computing instructors and a survey of 169, then summarize five recurring barriers: fragmented buy-in, policy crosswinds, implementation challenges, assessment misfit, and lack of resources. That combination makes the paper look like a descriptive, practice-level contribution with field-level implications. It is especially relevant to CHI because it broadens what counts as design work in AI-rich environments and highlights how commercial platform control can displace design responsibility onto instructors without giving them corresponding power. My main caution is evidentiary rather than conceptual. In the provided sections, nearly everything verifiable comes from the abstract, plus some contextual metadata and references. That is enough to confirm the headline framing, the mixed-method scope, and the stated barrier taxonomy, but not enough to inspect recruitment, coding, survey quality, or whether the recommendations are tightly supported by the data. So I would rate the paper as promising and likely influential as a framing reference, while keeping some methodological humility until the full body text is available for audit. Even with that caveat, the contribution appears timely, well scoped, and genuinely useful for future HCI work on educator-centered AI governance and classroom adaptation.

What Changed

Canon before

Prior HCI and computing-education work on GenAI has largely focused on student use, tool capabilities, classroom deployments, or instructor attitudes toward adoption. In that broader canon, instructors are often treated as policy setters, assignment designers, or adopters of educational tools. This paper shifts attention to a harder setting: instructors trying to influence student-AI interaction while the relevant interfaces are commercial systems outside instructor control. That framing connects AI-in-education concerns with HCI questions about indirect design, institutional constraints, and the practical labor of adaptation under uncertainty.

Departure from common sense

A common-sense reading of classroom technology change is that meaningful pedagogical design requires direct control over the tool, platform, or interface students use. This paper pushes against that assumption by arguing that instructors are still doing consequential design work even when they cannot modify the GenAI system itself. Their interventions are reactive and indirect, but still shape student-AI interaction through course redesign, assessment choices, and policy responses. The contribution is therefore not just that instructors face barriers, but that this constrained, improvised labor should itself be recognized as a design setting rather than dismissed as mere coping or compliance.

Actual novelty

The paper’s main novelty is a conceptual framing backed by mixed-method empirical evidence. It introduces the term “emergency pedagogical design” to name a recurring situation in which programming instructors must rapidly redesign coursework around generative AI despite lacking control over the commercial interfaces students use. The paper then operationalizes that framing through interviews and a survey, identifying five recurring barriers that structure this work: fragmented buy-in, policy crosswinds, implementation challenges, assessment misfit, and lack of resources. So the novelty is not a new artifact or intervention technique; it is a field-facing framework for understanding instructor labor under GenAI disruption, plus an empirically grounded barrier taxonomy that can orient future HCI and education research.

Evidence

The available focused sections still provide only a narrow evidentiary window, but they do support the paper’s headline contribution with exact text. The abstract explicitly defines “emergency pedagogical design” as reactive, indirect instructor work undertaken without control over commercial interfaces, which substantiates the paper’s departure from interface-centric assumptions about where design happens. The same abstract also states the empirical scope—interviews with 13 instructors and a survey of 169 computing instructors—and enumerates five recurring barriers. In addition, the metadata sections corroborate that the paper is positioned as an HCI empirical research article and list author tags such as “emergency pedagogical design,” “student-AI interaction,” “programming instructors,” and “generative AI.” What remains missing are methods, findings, discussion, and limitations sections, so the evidence can verify framing, scope, and topical positioning, but not analytic rigor or author-reflexive limitations in the usual depth.

Limits

Method limits

The available packet does not include the body sections where recruitment, interview protocol, survey construction, coding procedures, or analytic decisions would normally be described. As a result, the mixed-method design can be verified only at the level of stated sample sizes and claimed barrier themes, not at the level of methodological execution. That limits confidence about representativeness, triangulation quality, and how robustly the five barriers emerged across the two data sources. Any stronger judgment about rigor would require the missing methods and findings sections.

Deployment limits

The contribution is explicitly centered on programming instructors and on coursework shaped by commercial generative AI tools. That makes the framing most credible for higher-education computing contexts where instructors lack interface control yet must still respond to widespread student use. Transfer to non-programming disciplines, K–12 settings, or environments using institutionally controlled AI systems is plausible but not demonstrated in the provided text. The recommendations may therefore travel unevenly across educational contexts with different governance structures, assessment practices, and resource constraints.

Boundary conditions

The paper’s claims are bounded by a specific sociotechnical configuration: computing instructors, student use of commercial GenAI systems, and instructional settings where student-AI interactions are only partially visible to educators. The framing of emergency pedagogical design is most applicable when instructors must react quickly, cannot directly redesign the AI interface, and instead work through assignments, policies, and assessment structures. If an institution controls the AI tool, has stable governance, or can instrument student interaction directly, the paper’s central problem setting may change substantially. These conditions matter because the contribution is about constrained indirect design, not all forms of AI-supported teaching.

Manual check: Yes: the focused sections provide only abstract-level claims, metadata, and references, so the methods, findings, discussion, and any explicit author-stated limitations could not be directly verified from richer body sections.

Position in field

This paper sits in CHI as a conceptual and empirical framing contribution rather than a system, tool, or algorithm paper. It extends educator-centered HCI conversations about GenAI by naming a distinct design setting and by making instructor adaptation legible as design labor under platform and institutional constraint. Relative to adjacent work on AI in computing education, its value appears to be synthesis and reframing: it organizes a fast-moving practical problem into a vocabulary that HCI researchers, institutions, and support organizations can use. That gives it field-level relevance, especially for work on AI governance, educational infrastructure, and the limits of interface-centric intervention models.

Abstract

Generative AI (GenAI) tools are increasingly pervasive, pushing instructors to redesign how students use GenAI tools in coursework. We conceptualize this work as emergency pedagogical design: reactive, indirect efforts by instructors to shape student-AI interactions without control over commercial interfaces. To understand practices of lead users conducting emergency pedagogical design, we conducted interviews (n=13) and a survey (n=169) of computing instructors. These instructors repeatedly encountered five barriers: fragmented buy-in for revising courses; policy crosswinds from non-prescriptive institutional guidance; implementation challenges as instructors attempt interventions; assessment misfit as student-AI interactions are only partially visible to instructors; and lack of resources, including time, staffing, and paid tool access. We use these findings to present emergency pedagogical design as a distinct design setting for HCI and outline recommendations for HCI researchers, academic institutions, and organizations to effectively support instructors in adapting courses to GenAI.