Venus and Mars on Canvas: AI-Mediated Collaborative Drawing for Romantic Relationship Insight

Paper Title

Venus and Mars on Canvas: AI-Mediated Collaborative Drawing for Romantic Relationship Insight

Publication Info

• Topic area: AI-mediated tools for relationship development through collaborative creative activities.
• Keywords: AI-mediated collaboration, romantic relationships, self-reflection, partner acceptance, relational awareness, collaborative drawing, non-verbal communication, behavioral analysis, personalized insights, relationship technology, art therapy.

Background and Problem

• Problem / challenge: Traditional art therapy for couples is therapist-dependent, lacks systematic real-time behavioral analysis, and is often inaccessible due to logistical and financial barriers. Existing digital tools for couples focus on either creative expression or behavioral analysis in isolation, failing to integrate these elements for deeper relational insights.
• Significance: Addressing these gaps can provide couples with accessible, scalable tools to explore relational dynamics, foster self-reflection, and improve communication without requiring professional facilitation.
• Motivation and related work: Collaborative art therapy has shown therapeutic value in revealing relational patterns through non-verbal interactions. However, it remains limited by reliance on therapists and retrospective analysis. Digital tools like emotion recognition systems and VR co-experiences enhance emotional awareness but lack mechanisms for systematic behavioral analysis and personalized feedback. This paper seeks to combine the benefits of collaborative art with AI-driven analysis to create a comprehensive relationship support tool.

Solution

• Proposed approach: An AI-mediated collaborative drawing system that enables couples to engage in structured drawing activities while capturing and analyzing their interactions to provide personalized relational insights.
• Novelty:
  1. Integration of behavioral data collection, multimodal analysis, and AI-generated personalized questions and reports.
  2. Use of non-verbal communication through turn-based collaborative drawing to reveal relational dynamics.
  3. Development of a structured pipeline for real-time behavioral tracking and reflective feedback.
  4. Design considerations for AI-mediated tools that balance accessibility, engagement, and relational insight.
• Procedure and key techniques:
  • Couples engage in three themed drawing sessions (free drawing, date night, gift) using a shared digital canvas with turn-taking enforced.
  • Behavioral data (e.g., drawing duration, focus scores, area usage) is collected at turn, participant, and session levels.
  • AI generates personalized reflective questions based on observed behaviors and synthesizes a final report combining behavioral data, participant reflections, and drawings.
  • The system emphasizes non-verbal communication, playful engagement, and structured reflection to foster relational awareness.

Results

• Concrete findings:
  • High ratings for system usability and engagement: Perceived Usefulness (M = 5.58), Perceived Ease of Use (M = 6.04), Perceived Enjoyment (M = 6.50), and Intention to Use (M = 5.88).
  • Positive impact on relational outcomes: Self-reflection (M = 4.68), Partner Acceptance (M = 5.40), and Relational Awareness (M = 5.47).
  • High expert validation scores for AI-generated questions (M = 5.94) and reports (M = 6.14) across dimensions like relevance, reflection, and coherence.
• Advantage over baselines:
  • Combines creative expression and behavioral analysis, unlike existing tools that address these aspects separately.
  • Provides structured, evidence-based insights that are personalized and actionable, surpassing generic relationship advice.
  • Facilitates non-verbal communication and playful engagement, creating a unique relational exploration experience.
• Experiments / evaluation:
  • User study with 20 couples (N = 40) aged 18–32, with relationship durations ranging from 3 months to 9 years.
  • Mixed-method evaluation combining quantitative surveys (e.g., TAM, relational awareness) and qualitative interviews.
  • Expert validation of AI outputs by three professional art therapists, confirming appropriateness and quality.
• Limitations and future work:
  • Limited exploration of conflict resolution and challenging relational dynamics.
  • Current AI analysis focuses on behavioral patterns rather than deeper symbolic or metaphorical interpretations of drawings.
  • Study conducted in a specific cultural and age demographic; findings may not generalize to other populations.
  • Long-term impact on relationship dynamics remains untested; future work should include longitudinal studies and integration with professional resources.

Summary

This paper introduces an AI-mediated collaborative drawing system designed to enhance self-reflection, partner acceptance, and relational awareness in romantic relationships. By integrating creative expression with AI-driven behavioral analysis and personalized feedback, the system provides couples with accessible tools for exploring relational dynamics. A user study with 20 couples demonstrated high usability, engagement, and relational impact, while expert validation confirmed the quality of AI-generated outputs. The system’s emphasis on non-verbal communication and playful engagement offers a novel approach to relationship development, with potential for broader applications and future enhancements to address conflict resolution and long-term relational growth.