Research

I study how AI systems interact with human judgment, behaviour, relationships, and social institutions. Current work includes human–LLM moral judgment, recommender systems, artificial companionship, and AI-mediated communication.

I study cognitive offloading, attention and learning, including the efficiency–atrophy paradox: when technological assistance improves immediate performance while reducing opportunities to exercise the underlying capacity. Through Moral Red Lines, I examine limits on delegation, drawing on cognitive science, ethics, philosophy and theological anthropology.

I develop adaptive and approximate AI, on-device inference, and federated learning for resource-constrained systems. Applications include wearable and smartphone cardiac monitoring, UAV-based precision agriculture, and LLM-assisted OCR for cultural heritage.