Quantifying Co-Creative Sense-Making and Interaction Dynamics in Human–AI Collaboration
This research explores how co-creative interaction between humans and AI systems can be computationally modeled, quantified, and analyzed as a dynamic process of collaborative sense-making.
Published as an arXiv preprint, the paper introduces an experimental framework for studying human–AI co-creation not merely through final outputs, but through the evolving interactional dynamics that emerge during creative collaboration itself.
The paper presents the development of an interactive AI Drawing Partner and introduces the: Co-Creative Sense-Making (CCSM) Framework.
The CCSM framework provides a structured method for analyzing and quantifying co-creative interaction between humans and AI systems across multiple dimensions of collaboration and participation.
As AI systems become increasingly integrated into creative workflows, a fundamental challenge emerges:
How do we measure and understand the collaborative process unfolding between humans and AI systems during co-creation?
Traditional computational creativity research has often focused on: output quality, novelty, style, or autonomous generation.
This work instead focuses on: the interaction itself. The paper investigates how humans and AI systems: shape creative trajectories together, influence one another dynamically, regulate participation, and sustain collaborative interaction across time.
Rather than treating co-creativity as isolated generation, the work frames co-creative interaction as a process of collaborative sense-making emerging through reciprocal participation.
At the center of the paper is the development of an AI Drawing Partner — an interactive co-creative drawing system designed to participate dynamically within human creative workflows.
Rather than functioning as a passive generation tool, the system collaborates with the user through ongoing interaction, adapting to creative direction and participating within the evolving drawing process itself.
The AI Drawing Partner serves as an experimental platform for investigating: co-creative interaction, collaborative adaptation, interactional responsiveness, and human–AI creative dynamics.
This system helped establish an early foundation for later work in: co-creative AI, interaction-centered AI systems, participatory sense-making, and enactive approaches to Human–AI Interaction.
The paper introduces the CCSM framework as a way of organizing and analyzing co-creative interaction across multiple dimensions.
The framework investigates how collaboration evolves through:
How participants: generate ideas, shift conceptual direction, adapt creatively, and regulate exploration during collaboration.
How co-creative interaction unfolds through: responsiveness, turn-taking, contribution flow, and temporal interaction patterns.
How collaborative balance and participation emerge between: human contributors, AI systems, and distributed creative processes.
How co-creative interaction manifests within specific creative domains such as: drawing, visual art, ideation, and collaborative design.
Together, these dimensions create a framework for computationally studying co-creation as an evolving relational process.
A major contribution of the paper is the proposal that co-creative interaction itself can be analyzed as data.
Rather than evaluating only final artifacts, the work investigates how computational systems can model: collaborative adaptation, interactional balance, participation dynamics, creative influence, and evolving trajectories of shared sense-making.
This represents an important shift toward: interaction-centered approaches to computational creativity and Human–AI collaboration.
This paper serves as an important early foundation within the broader research ecosystem surrounding: Co-Creative AI Human–AI Interaction, Enactive AI, Participatory Sense-Making, Distributed Cognition, and Adaptive Regulation.
Many concepts explored in this work later evolved into broader frameworks including: Co-Creative Sense-Making, Participatory Coherence, Interactional Drift, Enactive Co-Creative AI, and Enactive Regulation Theory.
The paper represents an early bridge between: computational creativity, interactive AI systems, and enactive approaches to cognition and collaboration.
The AI Drawing Partner project contributes toward a broader vision of AI systems designed not merely to generate creative outputs, but to participate meaningfully within collaborative human creative processes.
Across this work, co-creativity is understood as: adaptive participation, collaborative sense-making, dynamic interaction, and the regulation of shared creative trajectories across time.
The long-term goal is to help establish frameworks for designing human–AI systems capable of sustaining meaningful co-creative interaction through adaptive collaboration, participatory intelligence, and interactional coherence.
2501.06607.pdf