Inclusion Analytics in Human-AI Collaborative Learning

A recent study published on arXiv presents a new approach to measuring inclusion in collaborative learning between humans and artificial intelligence. Traditionally, inclusion is assessed through sample descriptors or self-reports, missing the dynamics that unfold moment by moment.

A Discourse-Based Framework

The paper introduces "inclusion analytics," a discourse-based framework that examines inclusion as an interactive and dynamic process. This approach conceptualizes inclusion along three complementary dimensions: participation equity, affective climate, and epistemic equity. The goal is to make these dimensions analytically visible through scalable, interaction-level measures.

Applications and Results

Using both simulated conversations and empirical data from human-AI teaming experiments, the researchers demonstrate how inclusion analytics can surface patterns of participation, relational dynamics, and idea uptake that remain invisible to aggregate or post-hoc evaluations. This work represents an initial step toward process-oriented approaches to measuring inclusion in human-AI collaborative learning environments.