Quantified Co-Creative AI: 

Modeling, Explaining, and Visualizing Human–AI Creative Interaction

-Nicholas Davis, PhD


As AI systems become increasingly integrated into creative practice, a fundamental challenge emerges:

How do we understand what actually happens during co-creation between humans and AI systems?

Traditional generative AI systems often function as black boxes. While they may produce compelling outputs, the collaborative processes that shape those outputs — including interaction patterns, participation dynamics, creative influence, communication rhythms, and evolving creative trajectories — are frequently hidden within the final artifact itself.

Quantified Co-Creative AI explores a different approach.

Rather than evaluating only: final outputs, novelty, or stylistic quality, this research program investigates: the interaction dynamics of co-creation itself.

The central premise of Quantified Co-Creative AI is that co-creative interaction can be: modeled, quantified, visualized, analyzed, and explained computationally.

This work introduces frameworks and systems for studying how humans and AI systems collaboratively generate meaning, regulate participation, adapt creatively, and sustain interaction across time.

What Is Quantified Co-Creative AI?

Quantified Co-Creative AI refers to co-creative systems that computationally model and analyze the interaction dynamics unfolding during human–AI collaboration.

These systems do not simply generate creative artifacts.

Instead, they also capture: interaction histories, collaboration structures, communication patterns, turn-taking behavior, cognitive dynamics, and evolving creative trajectories.

This transforms co-creation itself into: analyzable interaction data.

Within this framework, creativity is treated not merely as a final product, but as a dynamic relational process emerging through participation and interaction.

Core Research Themes

Co-Creative Sense-Making

At the center of this work is the concept of: Co-Creative Sense-Making (CCSM)

CCSM investigates how humans and AI systems collaboratively construct meaning through interaction across time.

Rather than viewing creativity as isolated generation occurring independently within either participant, CCSM frames creativity as: participatory, adaptive, relational, and interaction-centered.

The framework analyzes how collaboration evolves through: cognitive dynamics, interaction dynamics, collaboration dynamics, and domain-specific creative behaviors.

Interaction Dynamics

A central contribution of this research program is the proposal that: interaction itself can be quantified.

This includes modeling: responsiveness, interaction rhythms, creative flow, communication patterns, collaborative balance, and shifts in participation.

By visualizing and analyzing interaction dynamics, quantified co-creative systems provide insight into: how collaboration unfolds, how ideas emerge, and how co-creative trajectories evolve over time.

Explainable Co-Creative AI

Quantified co-creative systems also function as: explainable co-creative systems.

Rather than explaining only isolated model outputs, these systems explain: interaction history, collaborative influence, creative provenance, participation structures, and the evolution of creative decisions.

This creates a framework for: transparent co-creation, explainable artificial media, and computational models of collaborative creativity.

Creative Provenance

This work proposes that interaction histories can function as: creative provenance.

By preserving the dynamics of collaboration, quantified co-creative systems make it possible to examine: who contributed what, how ideas evolved, when interaction shifted, and how collaboration shaped the final artifact.

This creates new possibilities for: explainable AI, collaborative authorship analysis, synthetic media transparency, and adaptive creative systems.

Adaptive Co-Creative Systems

Quantified interaction modeling also enables: adaptive participation.

By analyzing interaction dynamics across time, co-creative AI systems can: model collaborative preferences, adapt to user behavior, regulate participation, balance novelty and predictability, and sustain more meaningful long-term collaboration.

This creates pathways toward: adaptive co-creative agents, interaction-centered AI systems, and participatory artificial intelligence.

Research Areas

This section includes research across: Creative Sense-Making, AI Drawing Partner, Quantified Co-Creative AI Systems, Explainable Artificial Media, Interaction Dynamics Visualization, Participatory Coherence, Adaptive Co-Creative Systems, and Human–AI Creative Collaboration.

Together, these works contribute toward a broader research program investigating: how humans and AI systems sustain meaningful co-creative interaction through participation, adaptation, and collaborative sense-making.

Research Direction

Quantified Co-Creative AI represents a shift away from viewing AI creativity as isolated autonomous generation.

Instead, this research program investigates AI systems as:

Across this work, creativity is understood as: an emergent interactional process unfolding between humans and intelligent systems across time.

The long-term goal is to help establish a framework for: explainable co-creative AI, quantified collaboration systems, adaptive participatory AI, and interaction-centered approaches to Human–AI creativity.