Proxy Measures for Trust in Human-AI Interactions

This project explores how behavioral cues, such as politeness and follow-up questions, can act as proxies to measure trust and usefulness in human-AI interactions. By analyzing conversational patterns, we aim to develop scalable methods for evaluating the credibility and performance of AI systems without relying on intrusive surveys. Through controlled experiments and real-world data validation, this work seeks to bridge the gap between user perceptions and AI response quality, offering new insights into improving human-AI communication.