This research explores how Large Language Models can make AI systems that monitor daily activities in smart homes more explainable and transparent to users.

• Evaluates LLMs' ability to generate natural language explanations for activity recognition systems
• Compares different prompting techniques to enhance explanation quality
• Demonstrates how explanations help users understand which sensor data influenced AI decisions
• Identifies challenges and limitations in applying LLMs to IoT data interpretation

For healthcare applications, this approach enables more trustworthy monitoring systems for elderly care and patient independence assessment, potentially increasing adoption of smart home health technologies while respecting user privacy and autonomy.

Leveraging Large Language Models for Explainable Activity Recognition in Smart Homes: A Critical Evaluation