Paper Type

ERF

Abstract

Conversational AI tools are increasingly embedded in collaborative problem‑solving settings, yet we still know little about how they shape the distribution of cognitive work, coordination, and shared understanding within teams. Using a distributed‑cognition lens, we investigate how dyads engage in hybrid human–AI problem solving during a structured analytical task. A five‑step activity design (individual attempt, partner comparison, AI consultation, reconciliation, and reflection) enables us to trace how conversational AI contributes to representational repair, clarification, and coordination. Early observations show that AI’s collaborative role depends on whether teams begin with consistent or inconsistent initial attempts: in consistent cases, AI primarily validates and clarifies shared representations, whereas in inconsistent cases, AI provides a broader range of mediation functions—such as recomputation, strategy support, clarification, and reconciliation—to stabilize and align understanding. This work provides early insight into how conversational AI functions as a collaborative partner, informing future research and design.

Paper Number

1822

Comments

SIG HCI

Share

COinS
 
Aug 15th, 12:00 AM

Conversational AI as a Collaborative Partner: A Distributed Cognition Lens on Hybrid Human–AI Problem Solving

Conversational AI tools are increasingly embedded in collaborative problem‑solving settings, yet we still know little about how they shape the distribution of cognitive work, coordination, and shared understanding within teams. Using a distributed‑cognition lens, we investigate how dyads engage in hybrid human–AI problem solving during a structured analytical task. A five‑step activity design (individual attempt, partner comparison, AI consultation, reconciliation, and reflection) enables us to trace how conversational AI contributes to representational repair, clarification, and coordination. Early observations show that AI’s collaborative role depends on whether teams begin with consistent or inconsistent initial attempts: in consistent cases, AI primarily validates and clarifies shared representations, whereas in inconsistent cases, AI provides a broader range of mediation functions—such as recomputation, strategy support, clarification, and reconciliation—to stabilize and align understanding. This work provides early insight into how conversational AI functions as a collaborative partner, informing future research and design.

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.