I-Corps: Translation Potential of Privacy Preserving Semantic Artificial Intelligence (AI) Camera

Award ID: 2631505

This I-Corps project is based on the development of an artificial intelligence (AI) platform that provides continuous monitoring without capturing or storing images. Continuous monitoring protects older and vulnerable individuals in homes, assisted living facilities, and hospitals, where falls, prolonged inactivity, and other safety incidents often go undetected until it is too late. Existing vision-based monitoring systems raise privacy and dignity concerns, limiting their adoption in sensitive settings. This technology senses activity patterns and human-environment interactions at the point of capture, converting visual information into structured, non-visual context before any image is stored, eliminating sensitive data at its source rather than protecting it after the fact. The applications of this technology include elderly care, healthcare, smart home safety, workplace safety, and physical security, wherever monitoring is valuable, but privacy is a barrier to adoption. This technology may reduce regulatory and ethical risks in AI-enabled surveillance while improving safety outcomes for vulnerable populations. This I-Corps project utilizes experiential learning, coupled with firsthand investigation of the industry ecosystem, to assess the translation potential of a semantic artificial intelligence (AI) sensing platform that performs context recognition at the edge. Current systems use image capture and record raw video, which raises privacy concerns. This technology extracts structured semantic representations, such as posture transitions and activity patterns, using a domain-specific reasoning framework designed for context-aware inference. It embeds privacy-by-design directly into the sensing architecture, distinguishing it from existing camera-based systems that rely on post-hoc anonymization or encryption of retained images. The technology has been shown to provide reliable detection of safety-critical events, including falls and abnormal behavior, while maintaining zero raw-image retention. Users may benefit through continuous safety monitoring that preserves dignity and privacy, lowering barriers to adoption in regulated, safety-critical environments. 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.

NSF Program Director: Ruth Shuman