Career: modeling trust dynamics and system adaptation for human-AI collaboration in engineering design

The National Science Foundation Division of Civil, Mechanical, and Manufacturing Innovation awarded The Trustees of The Stevens Institute of Technology $545,650 on September 1, 2026, under the Faculty Early Career Development (CAREER) program to model trust dynamics and system adaptation for human-AI collaboration in engineering design.

The project develops empirical datasets and computational inference models that estimate trust as a continuous, time-varying state variable from synchronized behavioral, self-report, and psychophysiological indicators during real-time design interaction. The recipient will formalize trust-aware AI adaptation strategies that specify how AI feedback behavior should adjust to support calibrated reliance and effective human-AI collaboration. Education and outreach activities will deploy AI-assisted design tools in undergraduate engineering design courses and pre-college design bootcamps, integrating multimodal sensing to enable adaptive, personalized feedback for cognitively demanding learning. The research targets applications in high-stakes industrial domains including manufacturing, healthcare, and infrastructure, with the goal of improving engineering decision quality and reliability, reducing design cycle time and downstream rework.

Performance occurs in Hoboken, New Jersey with a period of performance from September 1, 2026, through August 31, 2031. This is assistance funding under the NSF Engineering program (CFDA 47.041).

Abstract

This faculty early career development program (career) project supports research and education focused on trust in human-artificial intelligence (AI) collaboration during early-stage engineering design. In open-ended, creative design contexts, trust governs how designers engage with and rely on AI-generated suggestions, yet this reliance is often implicit, context-dependent, and difficult to capture by traditional acceptance-based or post-hoc survey measures. This project advances fundamental knowledge by modeling trust as a continuous, time-varying latent state variable grounded in designers? Affective and cognitive processes, creating a foundation for trust-aware human-AI interaction in early-stage design. This career award advances theory and methods for trust in human-AI collaboration through two integrated research activities: (1) developing empirical datasets and computational inference models that estimate trust dynamics as a continuous, normalized quantity from synchronized behavioral, self-report, and psychophysiological indicators during real-time design interaction; and (2) formalizing trust-aware AI adaptation strategies that specify how AI feedback behavior should adjust to support calibrated reliance and effective collaboration. Education and outreach activities will deploy AI-assisted design tools in undergraduate engineering design courses and pre-college design bootcamps, and will integrate multimodal sensing to enable adaptive, personalized feedback that supports cognitively demanding learning. This research includes advancing the design and deployment of trust-aware AI systems that improve engineering decision quality and reliability, reduce design cycle time and downstream rework, and support effective human-AI collaboration across high-stakes industrial domains such as manufacturing, healthcare, and infrastructure. In parallel, the project will enhance stem education and workforce development by integrating AI-assisted design and personalized learning tools into undergraduate and pre-college programs, broadening participation and preparing an AI-literate engineering workforce. 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.- subawards are not planned for this award.