Opportunities and Barriers for AI Feedback on Meeting Inclusion in Socioorganizational Teams

Paper Title

Opportunities and Barriers for AI Feedback on Meeting Inclusion in Socioorganizational Teams

Publication Info

• Topic area: AI-mediated feedback systems for improving meeting inclusion in organizational contexts.
• Keywords: AI feedback, meeting inclusion, virtual meetings, Induced Hypocrisy Procedure, organizational barriers, sociotechnical systems, behavior change, human-AI interaction, feedback exchange, team dynamics.

Background and Problem

• Problem / challenge: Meetings often exclude certain voices, reducing their effectiveness. Feedback exchange can improve inclusion but is hindered by social dynamics, fear of confrontation, and organizational barriers.
• Significance: Inclusive meetings enhance creativity, productivity, and team cohesion, making them critical for organizational success.
• Motivation and related work: Prior research highlights the difficulty of giving and receiving feedback due to interpersonal and hierarchical challenges. Existing AI systems for meetings focus on awareness but struggle to induce behavior change. This paper explores whether AI agents can mediate feedback exchange and promote inclusion effectively.

Solution

• Proposed approach: An AI agent named Emily, designed to mediate feedback exchange and improve meeting inclusion using the Induced Hypocrisy Procedure (IHP), a psychological technique that highlights inconsistencies between values and behaviors to prompt change.
• Novelty:
  1. Application of IHP to AI-mediated feedback for meetings.
  2. Design of a sociotechnical system for facilitating inclusive behavior in virtual meetings.
  3. Empirical evaluation in both controlled and organizational contexts.
  4. Identification of organizational barriers to AI-mediated feedback adoption.
• Procedure and key techniques:
  • Post-meeting: Emily solicits feedback privately, anonymizes it, and routes it to intended recipients.
  • Pre-meeting: Emily guides users through IHP, asking them to set inclusion-related goals and reflect on past inconsistencies.
  • During meeting: Goals are displayed persistently, and speaking time data is collected for subsequent feedback.

Results

• Concrete findings:
  • Lab study (n = 28): Emily improved perceived meeting quality and balanced speaking times, with significant effects on meeting participation and efficiency.
  • Field study (n = 10): Emily influenced behavior but was primarily used for personal reflection due to organizational barriers.
• Advantage over baselines: Emily's use of IHP led to measurable behavior changes in the lab, outperforming prior systems that failed to induce such changes.
• Experiments / evaluation:
  • Lab study: Within-subjects design with structured tasks and questionnaires measuring meeting participation, quality, and group attraction.
  • Field study: Real-world deployment in a consulting firm, with quantitative surveys and qualitative interviews to assess adoption challenges.
• Limitations and future work:
  • Limited generalizability due to single organizational context.
  • Need for customization based on team dynamics and cultural factors.
  • Future work should address long-term behavior change and integrate systems into existing organizational structures.

Summary

This paper introduces Emily, an AI agent designed to improve meeting inclusion through feedback mediation and the Induced Hypocrisy Procedure. Lab results demonstrate Emily’s ability to influence participation and enhance meeting quality, while field results highlight organizational barriers such as time pressures, contextual misalignment, and hierarchical dynamics. The findings suggest that AI-mediated feedback systems can be effective in small collaborative groups but require careful integration into organizational workflows. This work contributes insights into designing AI systems for behavior change and meeting inclusion, emphasizing the importance of addressing sociotechnical and organizational factors.