The study, published in PLOS One, was grounded in the Theory of Stress Appraisal, a framework from psychology that describes how people evaluate demanding situations. According to this framework, when individuals face a stressor, they do not respond to the objective event but to their subjective reading of it. If a situation is interpreted as an opportunity for growth and mastery, it becomes a challenge appraisal; if it is interpreted as a threat that exceeds one’s resources, it becomes a hindrance appraisal. The researchers applied this lens to human-AI collaboration in design education, treating the collaboration itself as a stressor that students appraise in fundamentally different ways, with correspondingly different consequences for their developing competence.

The empirical investigation involved a questionnaire survey administered to design students across East China, with data analysis conducted using SPSS 24.0 and Amos 23.0. The analytical pipeline included reliability and validity testing to confirm the measurement instruments, regression analysis to estimate the relationships among variables, and Bootstrap mediation analysis to test the indirect pathways connecting collaboration to competence. The sample size of 483 provided sufficient statistical power to detect the effects the researchers hypothesized, and the combination of techniques allowed them to trace both the direct and indirect routes by which collaboration with AI shapes design competence.

The results revealed a dual-pathway mechanism that the authors characterize as one positive and one negative. On the constructive side, human-AI collaboration fostered challenge appraisal, with a standardized coefficient of 0.283, statistically significant at the 0.01 level. Challenge appraisal in turn positively predicted competence development, with a coefficient of 0.363, also significant at the 0.01 level. The indirect effect flowing through this positive pathway was 0.103, significant at the 0.01 level. In plain terms, students who experienced working with AI as an energizing challenge, a chance to stretch their abilities, tended to develop stronger design competence as a result.

On the destructive side, the same collaboration also triggered hindrance appraisal, with a coefficient of 0.271, significant at the 0.01 level. Hindrance appraisal negatively predicted competence development, with a coefficient of minus 0.229, significant at the 0.01 level, producing a significant indirect effect of minus 0.062. This means that students who perceived AI collaboration as an obstacle, a source of cognitive overload or anxiety about their own relevance, tended to see their competence development suffer. Because both pathways operated simultaneously and neither fully accounted for the total effect, the researchers describe the mechanism as partial mediation, a dialectical process in which the same experience simultaneously builds and erodes capability depending on individual cognitive appraisal.

The study then introduced a boundary condition designed to explain when the balance tips toward benefit rather than harm: teacher AI leadership. This construct refers to the capacity of educators to lead in AI-augmented classrooms, guiding students through the uncertainty that intelligent tools introduce. The analysis uncovered an asymmetric moderating effect that the authors consider among their most practically important findings. Teacher AI leadership significantly buffered the positive impact of collaboration on hindrance appraisal, with a coefficient of minus 0.112, significant at the 0.01 level. In other words, when teachers provided strong AI leadership, students were less likely to interpret AI collaboration as a threat, mitigating the cognitive load and anxiety that arise from technological uncertainty.

Strikingly, however, teacher AI leadership did not amplify the positive pathway. Its effect on the relationship between collaboration and challenge appraisal was essentially nil, with a coefficient of 0.001 that failed to reach statistical significance. The researchers interpret this asymmetry as evidence that the core value of teacher AI leadership lies in risk mitigation rather than creativity stimulation. Strong leadership can protect students from the anxiety and overload that intelligent tools provoke, but it cannot by itself manufacture the sense of challenge that drives competence growth. That positive appraisal appears to emerge from the collaboration experience itself, something leadership can safeguard but not directly produce.

The authors argue that these findings overturn the assumption that the impact of human-AI collaboration on design competence is linear. Instead, the effect is dialectical, contingent upon individual cognitive appraisal, and therefore impossible to optimize by simply increasing the amount of AI exposure in the curriculum. The same tool that pushes one student toward mastery can push another toward avoidance, and the psychological reading of the situation, not the tool, determines which direction prevails. This reframing positions cognitive appraisal as the central variable that AI-augmented design education must address if it hopes to deliver on its promise of intelligent synergy rather than mere tool assistance.

At the practical level, the researchers call for a role transformation in which educators shift from knowledge deliverers to cognitive facilitators in AI-augmented design education. By systematically developing AI instructional competencies and reforming assessment systems, educators can guide students toward establishing a healthy relationship with human-AI collaboration. The assessment reform component is particularly consequential, since evaluation practices that reward only human-authored output may inadvertently push students toward hindrance appraisals, while practices that recognize effective collaboration could support challenge appraisals. The goal, according to the study, is the synergistic advancement of technological empowerment and humanistic literacy, ensuring that as machines take on more of the design workload, students develop the competence to direct, critique, and elevate that work rather than feel displaced by it.

The study arrives at a moment when design programs worldwide are racing to integrate generative AI into studio practice, often without a clear account of the psychological mechanisms that determine whether integration helps or hurts. By demonstrating that the outcome depends on appraisal, and that teacher leadership functions as a brake on harm rather than an accelerator of benefit, the research offers a theoretical foundation for the intelligent transformation of educational design. The message for institutions is measured: invest in teacher AI leadership to reduce the anxiety and cognitive load that technological uncertainty creates, but recognize that stimulating the growth-oriented challenge appraisal students need for competence development requires deeper changes in pedagogy, assessment, and the culture of the studio itself.

Subject of Research: The double-edged effect of human-AI collaboration on design students' competence development through challenge and hindrance appraisals

Article Title: The double-edged sword of human-AI collaboration on design students’ competence development: The mediating roles of challenge and hindrance appraisals and the asymmetric moderation of teacher AI leadership

Article References: Wang, X., Li, M., Du, J., Yao, Y., & Jiang, J. (2026). The double-edged sword of human-AI collaboration on design students’ competence development: The mediating roles of challenge and hindrance appraisals and the asymmetric moderation of teacher AI leadership. PLOS One, 21(10), e0360722. https://doi.org/10.1371/journal.pone.0360722

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0360722

Keywords: human-AI collaboration, design education, design competence, challenge appraisal, hindrance appraisal, teacher AI leadership, stress appraisal theory, generative AI, PLOS One, competence development, cognitive appraisal, East China