AI as an Agent and Collaborative Space: Exploring the role of Generative AI in Small Group Synchronous and Asynchronous Collaborative Dynamics

Paper Title

AI as an Agent and Collaborative Space: Exploring the role of Generative AI in Small Group Synchronous and Asynchronous Collaborative Dynamics

Publication Info

• Topic area: Generative AI's impact on collaborative dynamics in small-group teamwork.
• Keywords: Generative AI, collaboration, distributed cognition, synchronous teamwork, asynchronous teamwork, role negotiation, decision-making, shared space, human–AI interaction, CSCW.

Background and Problem

• Problem / challenge: Current applications of Generative AI (GenAI) are predominantly designed for individual use, with limited exploration of its role in multi-party collaboration. Existing studies lack a systematic understanding of how GenAI affects collaborative dynamics such as information flow, role negotiation, and decision-making.
• Significance: Understanding GenAI’s role in collaboration is critical for designing systems that enhance teamwork efficiency, transparency, and mutual awareness, especially as GenAI increasingly integrates into professional and educational contexts.
• Motivation and related work: Prior studies have explored GenAI’s potential in scaffolding discussions, supporting ideation, and improving learning outcomes in small-group settings. However, these studies often focus on short-term prototypes and fail to address the continuous influence of GenAI on evolving collaborative dynamics. This paper addresses these gaps using the theoretical framework of Distributed Cognition (DC).

Solution

• Proposed approach: GenAI-Supported Cooperative Work (GSCW) framework, which bridges Human–AI Interaction (HAII) and Computer-Supported Cooperative Work (CSCW) by analyzing GenAI’s role as both a configurable agent and a collaborative space.
• Novelty:
  1. Empirical insights into GenAI’s influence on collaborative dynamics, including information flow, role negotiation, and decision-making.
  2. Identification of GenAI as a shared collaborative space that anchors attention and preserves group memory.
  3. Design considerations for future GenAI systems to support both synchronous and asynchronous collaboration.
• Procedure and key techniques:
  • Conducted a qualitative study with 27 higher education students, including observations of synchronous teamwork (six groups using ChatGPT-4) and semi-structured interviews about asynchronous collaboration.
  • Analyzed collaborative dynamics using thematic coding informed by Distributed Cognition, focusing on information flow, role negotiation, decision-making, and GenAI’s role as a collaborative space.

Results

• Concrete findings:
  • In synchronous settings, GenAI supported transparency, mutual awareness, and served as a shared memory and attentional anchor.
  • In asynchronous settings, GenAI was used individually, with outputs filtered and introduced into group discussions, reducing opportunities for negotiation and transparency.
  • Teams exhibited three patterns of decision-making with GenAI: evaluated adoption, critical rejection, and overreliance.
• Advantage over baselines: The study highlights GenAI’s potential to act as a collaborative space and negotiable team member, surpassing traditional individual-centric AI tools by fostering shared awareness and group memory in synchronous settings.
• Experiments / evaluation:
  • Observations involved 18 participants divided into six groups, using a shared GenAI interface to collaboratively design a cultural educational game.
  • Interviews with nine participants explored broader asynchronous GenAI usage scenarios.
  • Data included 5 hours and 26 minutes of video recordings, 6 hours and 12 minutes of interview audio, and thematic coding of collaborative dynamics.
• Limitations and future work:
  • Limited participant diversity and focus on university students may restrict generalizability.
  • Observations were confined to text-based GenAI systems and short-term tasks.
  • Future research should explore diverse team types, multimodal AI tools, and longer-term deployments in real-world settings.

Summary

This paper investigates how Generative AI (GenAI) influences collaborative dynamics in small-group teamwork, focusing on information flow, role negotiation, and decision-making. Through observations and interviews, the study reveals that GenAI acts as a shared collaborative space in synchronous settings, enhancing transparency and mutual awareness, but is primarily treated as a private assistant in asynchronous contexts, limiting collective negotiation. The authors propose a GenAI-Supported Cooperative Work (GSCW) framework to bridge Human–AI Interaction and CSCW, emphasizing design considerations for future systems that integrate shared and private access while supporting group memory and decision-making. These findings contribute to understanding GenAI’s evolving role in teamwork and inform the design of collaborative AI systems.