An AI Co-Creation tool that generates creative, customized stories based on user input
for
Computational Creativity
===========
For the final Computational Creativity assignment, we developed PlotCraft, an interactive co-creative storytelling system that demonstrates human-AI collaboration when generating a story. I was responsible for the UI/UX design, user experience research and conducting user studies, while my teammate handled the technical implementation and ChatGPT integration. The challenge was to create a system that didn't just assist creativity but actively participated as a creative partner, exploring the boundaries between human imagination and AI storytelling capabilities.
We designed PlotCraft with a restaurant-themed interface where users adjust sliders to control genre intensities (horror, comedy, romance, etc.) and select artifacts to be woven into their stories. The system featured a virtual butler guide and generated stories paragraph by paragraph, allowing users to iteratively modify their choices and see immediate narrative consequences. Using ChatGPT as our AI collaborator, we developed specific prompts to transform user selections into cohesive story elements, with the AI balancing genre requirements while integrating unexpected twists and maintaining story consistency.
The most interesting aspect was our extensive brainstorming collaboration with ChatGPT during the design phase. We treated the AI as a genuine creative partner, exploring co-creativity concepts, generating story prompts, and even developing character customization options together. This meta-collaboration, using AI to design an AI co-creative system, provided deep insights into computational creativity and human-AI creative partnerships.
We evaluated PlotCraft through user studies with four Media Technology students, focusing on both the creative experience and the narrative output quality. Users consistently rated the system 4-5/5 for inspiring creativity, particularly praising the unconventional genre combinations and artifacts that pushed them to explore unexpected narrative territories. The evaluation revealed that successful co-creativity emerges from providing users with curated limitations rather than complete freedom - the specific genre and artifact options sparked more creativity than open-ended prompts would have. This project taught me that effective co-creative systems should surprise and challenge users while maintaining enough structure to produce coherent outputs, and that the most valuable AI collaboration often comes from embracing unexpected combinations rather than trying to replicate human creative processes.
﹌﹌﹌﹌﹌﹌﹌
PlotCraft demo
Files for this project
Research
Implementation
Evaluation