Collaborative Research: EAGER: Mitigating Technostress in AI-Mediated Undergraduate Engineering Labs

Artificial intelligence (AI) is rapidly influencing engineering education and the requirements for the future STEM workforce. Engineering education degree programs build engineering laboratories, in which students must integrate conceptual knowledge, procedural skills, and real-time decision-making. AI-supported instruction adds another layer of complexity: as engineering degree programs increasingly incorporate AI-driven tools, students are also expected to increasingly interpret automated feedback, manage complex and AI-influenced technological interfaces, and make rapid decisions in dynamic learning environments where AI is a factor. These complex demands can lead to technostress, an interdisciplinary and multidimensional concept that dynamically unfolds during task completion and is therefore difficult to investigate using conventional methods. Technostress is associated with reduced learning agency and diminished performance and persistence, potentially reducing students' learning potential when engaging with emerging technologies like AI, even when learning *about* emerging technologies like AI. This project will therefore address a fundamental challenge in the professional formation of engineers: how to prepare students to effectively operate in technology-rich environments while maintaining adaptive learning behaviors and students' well-being. This project will use exploratory methods whose scalability is uncertain to investigate an emerging phenomenon with little research base in this field -- that of technostress generated from using AI-enabled technologies in undergraduate learning – to establish foundational evidence for future research and educational innovation. By advancing the understanding of how technostress emerges during AI-supported learning and how it can be mitigated, the project will promote the progress of science at the intersection of engineering education, cognitive science, and human-centered AI. The project will advance NSF's mission by generating foundational knowledge that supports human-centered AI, strengthens the future STEM workforce, and improves the education of engineers and other technical workforces in increasingly automated and technology-rich workplaces. The findings will contribute to more effective engineering learning environments, support student success and persistence, and help prepare a larger, more resilient U.S. technical workforce for AI-enabled professions. This project will explore the feasibility of designing, implementing, and evaluating AI-supported scaffolding that enhances self-regulation in undergraduate engineering laboratories using multimodal physiological indicators of technostress. The project will: (1) identify the feasible indicators of technostress in real time during AI-mediated laboratory activities; (2) design and implement AI-based scaffolds that support students’ self-regulated learning processes; and (3) evaluate the feasibility and scalability of multimodal approaches for detecting technostress across engineering laboratory environments. The research will employ a multi-institutional, multimodal design involving undergraduate engineering students across varied laboratory contexts. The team will collect behavioral, self-reported, and physiological data (including skin conductance, eye tracking, and biological stress markers) from undergraduate engineering students at two universities to examine how technostress develops during laboratory learning. The team will align synchronize the timing of these two data streams with phases of self-regulated learning (such as planning, monitoring, and reflection) to map how technostress develops and changes during laboratory tasks. The project will compare data from equipment with different sensing capabilities drawn from different laboratory environments to determine which approaches provide sufficient information for scalable technostress detection. The team will use an AI-tool agnostic approach, meaning that the study design is more robust for handling differences in tool version or design over the course of the study. Expected outcomes include a conceptual framework linking technostress, self-regulation, engagement, and AI-supported learning; foundational evidence regarding the feasibility of multimodal technostress detection; and practical guidelines for designing adaptive, human-centered AI learning environments. Together, these contributions will advance engineering education, cognitive science, and human-centered AI while providing the scientific foundation for future adaptive learning environments that improve student learning, resilience, and workforce preparation. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.