Large language models reshape emotional tracking but have key limits
Scoping the Landscape of Large Language Model-based Affective Personal Informatics: Review, Reflection, Future Direction
Human-Computer Interaction
Summary
Tracking and understanding personal feelings can help people with their mental health. The authors looked at 45 studies using big language models (LLMs) to do this kind of emotional tracking, but found most approaches rely heavily on text and require active screen use. They also found that the studies mostly involved young people on university campuses, limiting broader relevance. The authors point out these problems and suggest future work should include more types of data, make systems easier to use, and include more diverse users.
What this means in practice
- •For mental health app developers: Improve emotional self-awareness tools by integrating multimodal data beyond text to reduce user burden and broaden user inclusion.
- •For human-computer interaction designers: Design affective informatics systems with less screen dependency and incorporate diverse user groups to enhance relevance and usability.
A survey. It maps existing work.
Authors
Minjie Liu, Lihuan Li, Xuran Li, Flora Salim, Benjamin Tag
Abstract
Affective personal informatics (PI) involves capturing, tracking, analyzing, and reflecting on personal data to support emotional self-awareness and mental well-being. Recent advances in large language models (LLMs) are reshaping affective PI, yet a comprehensive survey of LLM-based affective PI systems is lacking. To bridge this gap, we systematically analyzed 45 papers published between January 2020 and June 2026 across four major computing and health databases, examining their macro trends, system taxonomies, and research methodologies. Our synthesis reveals an overreliance on multi-stage, text-centric pipelines, a predominance of active, screen-bound interactions, and a severe demographic skew toward young, university-campus-based cohorts in empirical evaluations. Based on these findings, we critically reflect on prevailing architectural and methodological shortcomings and outline actionable future directions for building end-to-end multimodal, low-burden, and inclusive affective PI systems.