This article presents a conceptual framework for analyzing dialogue dynamics in collaborative problem-solving contexts, with a specific focus on human-AI and multi-agent interactions. The authors argue that understanding these dialogic interactions is crucial for optimizing partnerships as intelligent systems gain autonomous reasoning capabilities.
- The framework utilizes a hierarchical two-layer coding scheme that integrates cognitive and non-cognitive problem solving with metacognitive regulatory mechanisms.
- It addresses limitations in current analytical approaches by providing a structured method to evaluate collaborative partnerships.
- The authors demonstrate the framework's effectiveness and generalizability across nine datasets spanning multiple domains.
- Analysis reveals that metacognitive regulation serves as an essential discriminator for identifying deeper levels of collaboration between humans and agents.
The study provides insights into how humans and agents coordinate their knowledge, skills, and efforts to solve complex problems, highlighting the importance of metacognitive processes in effective collaborative problem solving.