Researchers have introduced Trajectory-based Disclosure Evaluation (TraDE), a new framework designed to better assess conversational AI's ability to handle evolving disclosure in human-AI interactions. Grounded in Social Penetration Theory, TraDE evaluates how a model's responses are shaped by the history of disclosure, not just the current conversational state. Experiments using TraDE revealed that models' reactions to disclosures can vary based on preceding conversational trajectories, particularly in how they accept or challenge user framing, and that these effects can persist even after disclosure levels return to shallow. The framework also highlighted that models often diverge from human supporters most notably when users reduce their level of disclosure. AI
IMPACT This framework could lead to more nuanced evaluations of AI's social and emotional support capabilities, potentially improving their relational behavior.
RANK_REASON The cluster contains an academic paper detailing a new evaluation framework for conversational AI. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →