IdeasLand: Sculpting Creativity with Constraints in Human-AI Collaboration An AI-first approach

Abstract

The Constrained Collaborative Environment (CCE) model introduces a structured framework for human-AI collaboration, leveraging predefined constraints to enhance creative ideation while preserving human agency. This paper evaluates CCE’s implementation through IdeasLand, a web application built with Golang, NuxtJS, and PostgreSQL, deployed on Google Cloud. A mixed-methods study with five participants in September 2024 assessed usability (SUS score: 72), creativity impacts (ratings: 3.6–3.8), and user experience, revealing strengths in AI-driven idea structuring and refinement, alongside challenges in interface reliability (e.g., “Journey” section errors) and design (e.g., dark palette critiques). The CCE’s Logic, Interface, and Instructions components effectively scaffold ideation, outperforming flexible models like MICI in efficiency, yet technical and usability gaps highlight refinement needs. Advanced features like AI-generated avatars and the Hero’s Journey tool showcase CCE’s potential, while qualitative feedback calls for brighter aesthetics and robust AI delivery. Future directions propose technical optimizations (e.g., real-time AI), interdisciplinary adaptability, and longitudinal validation to solidify CCE’s role in creative industries. This work contributes to HCI by demonstrating how structured AI partnerships can accelerate creativity, offering a scalable testbed for human-AI interaction design.