Sustainable Human-AI Collaboration in Creative Contexts: An Integrated Approach of CASA, UTAUT, Psychological Ownership, and Self-Determination Theory

Paper Title

Sustainable Human-AI Collaboration in Creative Contexts: An Integrated Approach of CASA, UTAUT, Psychological Ownership, and Self-Determination Theory

Publication Info

• Topic area: Psychological mechanisms in sustainable human–AI collaboration for creative tasks.
• Keywords: Human–AI collaboration, generative AI, psychological ownership, UTAUT, self-determination theory, CASA paradigm, creative contexts, collaboration satisfaction, performance expectancy, autonomy.

Background and Problem

• Problem / challenge: Existing frameworks like UTAUT inadequately address emotional and psychological dimensions (e.g., ownership, autonomy) in human–AI collaboration. Additionally, the impact of creative contexts (pure vs. work-related) and collaboration types (human-led vs. AI-led) on user perceptions and behaviors remains underexplored.
• Significance: Understanding these mechanisms is critical for fostering sustainable and satisfying human–AI collaboration, particularly as generative AI becomes integral to creative industries.
• Motivation and related work: Prior studies have focused on technology acceptance and satisfaction but overlooked deeper psychological factors like ownership and autonomy. This study builds on the CASA paradigm, UTAUT, and self-determination theory to explore these gaps in creative collaboration contexts.

Solution

• Proposed approach: Integration of UTAUT, psychological ownership theory, and self-determination theory to examine how AI performance expectancy influences psychological ownership, collaboration satisfaction, and continuance intention across creative contexts and collaboration types.
• Novelty:
  1. Extends UTAUT by incorporating psychological ownership and autonomy into human–AI collaboration.
  2. Empirically tests the CASA paradigm in creative contexts, treating AI as a social collaborator.
  3. Investigates moderating effects of creative contexts (pure vs. work-related) and collaboration types (human-led vs. AI-led).
  4. Proposes a multi-theory framework for sustainable AI collaboration in creative tasks.
• Procedure and key techniques:
  • Conducted a 2×2 experimental design (creative context: pure vs. work-related; collaboration type: human-led vs. AI-led) with 280 participants experienced in generative AI.
  • Measured AI performance expectancy, psychological ownership, collaboration satisfaction, and continuance intention using validated scales.
  • Analyzed data using Partial Least Squares Structural Equation Modeling (PLS-SEM) and Multi-Group Analysis (MGA) to test hypotheses and moderating effects.

Results

• Concrete findings:
  • AI performance expectancy significantly increased psychological ownership (β = 0.41, p < .001), collaboration satisfaction (β = 0.64, p < .001), and continuance intention (β = 0.16, p < .1).
  • Collaboration satisfaction strongly predicted continuance intention (β = 0.45, p < .001).
  • Psychological ownership partially mediated the relationship between performance expectancy and continuance intention (β = 0.04, p > .05), while satisfaction fully mediated this relationship (β = 0.30, p < .001).
• Advantage over baselines:
  • Demonstrated that psychological ownership and autonomy are critical drivers of sustainable AI collaboration, extending beyond traditional UTAUT models.
  • Found that AI-led collaboration weakens the positive effect of performance expectancy on satisfaction (β = −0.16, p < .1), highlighting the importance of user autonomy.
• Experiments / evaluation:
  • Sample: 280 participants (50% male, 50% female; aged 20–40) with prior generative AI experience.
  • Design: 2×2 factorial design with human-led vs. AI-led collaboration and pure vs. work-related creative contexts.
  • Metrics: AI performance expectancy, psychological ownership, collaboration satisfaction, continuance intention, and psychological reactance.
• Limitations and future work:
  • Relied on vignette-based scenarios rather than real-time collaboration, limiting ecological validity.
  • Excluded professional creators, whose workflows and concerns (e.g., copyright) may differ.
  • Conceptualized collaboration type dichotomously (human-led vs. AI-led), overlooking nuanced interaction dynamics.
  • Future research should explore interactive, task-based designs, include professional creators, and examine stage-specific AI roles in creative processes.

Summary

This study integrates UTAUT, psychological ownership theory, and self-determination theory to investigate sustainable human–AI collaboration in creative tasks. It finds that AI performance expectancy enhances psychological ownership, collaboration satisfaction, and continuance intention, with satisfaction serving as the strongest mediator. While creative contexts (pure vs. work-related) did not significantly alter these mechanisms, AI-led collaboration weakened satisfaction due to reduced user autonomy. These findings highlight the importance of designing AI tools that balance autonomy and performance while fostering ownership and satisfaction. Future research should adopt interactive designs and include professional creators to further refine these insights.