I-Corps: Translation Potential of a Privacy-Aware Stress Sensing and Conversational Companion

This I-Corps project is based on the development of a privacy-aware, on-device stress sensing and conversational artificial intelligence (AI) system. Current digital health tools present privacy and data misuse concerns that prevents patients from accessing continuous mental health monitoring. This technology delivers real-time, actionable coping interventions directly to the user through an AI conversational companion that keeps all sensitive physiological and behavioral data securely locked on the individual's personal device. By removing the liability and security risks associated with centralized data collection, this technology has application in the enterprise human resources and telehealth markets. This may provide proactive, secure mental health support, and reduce long-term healthcare costs without compromising privacy. This I-Corps project utilizes experiential learning coupled with first-hand investigation of the industry ecosystem to assess the translation potential of a multimodal stress inference engine and federated learning framework. This technology deploys a localized neural network model that aligns time-series wearable signals, such as heart rate variability and electrodermal activity, with mobile behavioral data to infer stress states in real time. Unlike conventional cloud-based wellness applications that centralize raw health data, this system achieves model personalization through adaptive, on-device fine-tuning against user-specific baselines. Global model improvements are coordinated via federated learning, securely aggregating encrypted weight updates without sharing raw data. Previous research demonstrated the feasibility of running robust stress inference prototypes strictly at the edge, offering a novel architectural paradigm for compliant, privacy-first digital healthcare. Users may benefit from a conversational tool that adjusts dialogue as a companion based on these localized inferences while strictly preserving privacy constraints. 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.