Volume 65

Large Language Models in Home Energy and Comfort Management: Opportunities, Framework, and Challenges Jiajing GAO1, Changying XIANG1*

Abstract

Recent advancements in Large Language Models (LLMs) have reshaped human-AI interaction across various domains, yet their application in Home Energy and Comfort Management (HECM) remains in its infancy. By extracting insights from recent studies integrating GPT-based agents, semantic sensors, and LLM-driven virtual datasets, we identify both emerging opportunities and unresolved challenges. The proposed analytical framework illustrates that LLMs can act as (1) semantic interfaces translating natural language into executable energy control commands; (2) reasoning engines enabling interpretable multi-objective optimization; (3) synthetic data generators supporting privacy-aware model training; and (4) adaptive mediators fostering human-centered energy behavior understanding. However, concerns remain regarding hallucination, security, local deployment, and model validation. Future research should integrate LLM reasoning with edge intelligence, federated learning, and human-in-the-loop feedback to achieve trustworthy, personalized, and sustainable HECM systems.

Keywords Large Language Model, Smart Home, Home Energy and Comfort Management