Meta-Self: Wearables and AI Systems for Augmenting Metacognition and Self-Regulation
Name
Kim-minsol-SM-MAS-2026-thesis.pdf
Size
7.53 MB
Format
Adobe PDF
Checksum (MD5)
0ba1437b83f3bc9bfad83e9a53e5f6d1
Author(s)
Kim, Minsol Michelle
Advisor(s)
Maes, Pattie
Date Issued
May 2026
Publisher
Massachusetts Institute of Technology
Abstract
Augmenting awareness of bodily states, emotional experiences, and behavioral patterns is not sufficient for effective self-regulation. Wearables can monitor physiological signals, conversational agents can elicit everyday reflections, and artificial intelligence (AI) can surface patterns or suggest alternatives, but none of these guarantees transformation. People often remain caught in repetitive negative cycles: aware of their struggles, yet unable to act differently. This thesis investigates how technology can cultivate MetaSelf: the human capacity for metacognition and self-regulation across everyday contexts. To this end, I design, deploy, and empirically evaluate human-centered MetaSelf Technologies (MSTs) that support MetaSelf skill development across the interrelated layers of body, emotion, and behavior. At the Body–Emotion layer, I examine how everyday usage of a watch providing real-time heartbeat-synchronized haptic feedback makes cardiac signals more perceptible for interoception training and affects emotion regulation and mental health outcomes. At the Emotion–Behavior layer, I introduce a Reflection-to-Action framework in which voice journaling captures emotionally salient, goal-misaligned experiences and uses these grounded instances to help individuals reflect on and decode their coping behaviors and emotional responses, engage in counterfactual “what-if” thinking, and generate concrete alternative action plans to revise future behavior. At the Behavior–Cognition layer, I study how human–AI interaction can expand counterfactual thinking and support decisions about future actions, evaluated through perceived utility and AI adoption behavior. Finally, I introduce the Reflective Agency Framework, grounded in phenomenology and Aristotelian ethics, to examine how wearable and AI systems can augment MetaSelf without displacing human agency. Drawing from behavioral psychology, affective science, and philosophy, this thesis contributes intervention systems, empirical evidence, and a design space for Human-Computer Interaction research on wearable and AI systems that cultivate metacognition and self-regulation toward self-directed change and sustained well-being.
MIT Department
Program in Media Arts and Sciences (Massachusetts Institute of Technology)
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link