Enactive AI: A Paradigm for Co-Creative AI and Creative Technologies

Nicholas Davis, PhD

Executive Summary

This research investigates how principles from enactive cognitive science can inform the design of next-generation interactive AI systems for creative practice and dynamic environments. Rather than treating intelligence as optimization over static data distributions, this work explores how intelligent systems can sustain coherent behavior through ongoing interaction with evolving contexts. To study these suppositions, I develop experimental creative technology platforms—including co-creative drawing agents, adaptive decision architectures, and instrumented creative interfaces—that allow the dynamics of human–AI collaboration to be measured and analyzed. These systems provide empirical testbeds for theories of participatory sense-making, enabling the study of how creativity, coordination, and meaning emerge through interaction between humans and AI. Together, this research program advances a paradigm of Enactive AI, in which intelligent systems support adaptive sense-making in creative collaboration, therapeutic creative practices, and other dynamic environments.

1. Introduction: Enaction as a Framework for Co-Creative AI

My research investigates how theories of sense-making from enactive cognitive science can inform the design of next-generation interactive AI systems for creative practice and dynamic environments. Many contemporary AI systems are designed to optimize prediction under relatively stable statistical assumptions. However, the environments in which people create, collaborate, and perform—from artistic practice and interactive media to human–AI creative collaboration—are inherently dynamic, with shifting relationships, evolving intentions, and continuously changing structural conditions.

My work explores how intelligent systems can participate meaningfully within these contexts by regulating their interaction with the environment rather than optimizing static objectives. Instead of treating intelligence and co-creation as improved prediction, this research investigates how AI systems can sustain shared sense-making during ongoing interaction, adapting their behavior as creative and environmental conditions unfold. To study these questions empirically, I develop interactive computational platforms that allow human–AI creative interaction to be observed, modeled, and analyzed in real time.

Central to this work is the concept of participatory sense-making, which describes how meaning emerges through interaction between agents as they coordinate perception and action within a shared environment. In creative collaboration, sense-making does not occur within an individual participant alone but unfolds through evolving interaction dynamics between collaborators. This research examines how participatory sense-making emerges during human–AI creative interaction, including how collaborators interpret evolving structures, coordinate actions in context, adapt to one another over time, and respond to emerging opportunities within the developing work.

This perspective frames creativity as a dynamic interaction process rather than a sequence of isolated outputs. By integrating principles of enactive cognition into interactive AI systems, this research explores how intelligent technologies can participate in the evolving structure of creative work. The goal is to develop AI architectures that do not simply generate artifacts but instead sustain meaningful collaboration by adapting to shifting creative contexts, human intentions, and emerging structures within the creative artifact itself.

2. Arts Practice and Origin of the Research

Alongside my research in co-creative AI and interactive systems, I maintain an active visual art practice centered on abstract painting and drawing, often created collaboratively with both novice and expert participants. Through this practice, I became interested in how artists perceive and select creative affordances during the act of drawing. My work explores how individuals navigate large spaces of possible artistic actions by attending to a subset of salient opportunities that emerge within the evolving visual structure. This process enables participants to engage with complex composition through intuitive interaction, and because the work is abstract—without predetermined correct outcomes—it lowers barriers to entry and allows novices to participate meaningfully alongside experienced artists. I also developed a drawing technique, the beckoning line, in which an initial stroke acts as a generative seed and subsequent marks emerge through local relationships within the evolving composition. This artistic practice serves as both a source of creative inquiry and a living laboratory for the study of participatory sense-making, informing the design of the co-creative systems developed throughout my research program.

2. Research Program Overview

My research investigates how creative technologies and intelligent systems can participate in collaborative creative processes with humans. I design computational systems that support co-creation between artists, designers, and artificial intelligence, with a particular focus on how creative collaboration unfolds through real-time interaction. Rather than treating AI as an autonomous generator of artifacts, my work explores how intelligent systems can function as creative partners that respond to human input, adapt to evolving creative contexts, and sustain meaningful interaction throughout the creative process.

I have been working in the field of co-creative AI since my early research helped lay conceptual foundations for human–computer co-creativity. In 2013, I published Human–Computer Co-Creativity: Blending Human and Computational Creativity, which explored how humans and computational systems can participate together in shared creative processes rather than treating AI as isolated generators of artifacts. This work contributed to a shift in how researchers conceptualize creative AI systems, emphasizing interaction, collaboration, and shared creative agency between human and machine participants.

Since then, my research has continued to develop theoretical frameworks and interactive systems for studying co-creative interaction. Through work on co-creative drawing agents and adaptive creative technologies, I investigate how intelligent systems can participate meaningfully in evolving creative practices. Taken together, this research program advances a paradigm of Enactive AI, in which artificial systems sustain adaptive sense-making across creative collaboration, therapeutic practice, and dynamic decision environments.

3. Co-Creative and Adaptive AI Research Platforms

This research program is guided by three central questions:

To investigate these questions, I develop computational platforms that function as technical probes for studying human–AI interaction.

3.1. Co-Creative Drawing Systems

My early work explored co-creative drawing systems that collaborate with users during the artistic process itself. One of the first prototypes, Drawing Apprentice, enabled real-time interaction between a human artist and a computational drawing partner. The system analyzed user drawings, responded with contextually relevant contributions, and learned through feedback provided during the drawing process.

This work evolved into the AI Drawing Partner, which functions both as a creative collaborator and as a research platform for studying human–AI interaction dynamics. The system captures detailed traces of the creative process—including cognitive, interactional, collaborative, and behavioral dynamics—allowing co-creation to be analyzed computationally. These interaction traces enable researchers to study how artistic meaning emerges through temporally extended collaboration between human and machine participants.

3.2. Aether: Enactive Co-Creative Drawing AI

Building on this work, I developed Aether, an enactive co-creative drawing agent designed to explore participatory sense-making in creative interaction. Unlike earlier systems that respond only to individual strokes, Aether evaluates the evolving structure of the artwork and the interaction dynamics between collaborators. The system regulates when to intervene, how much to contribute, and where to place marks on the canvas based on the developing visual composition. Aether integrates theoretical models developed in my research, including Creative Sense-Making, the Enactive Model of Creativity, and the Enactive Model of Consciousness, which conceptualize cognition as an emergent process arising from the regulation of perception, action, and environmental coupling. Through these mechanisms, Aether functions as both a creative partner and an experimental platform for studying how humans and artificial agents coordinate perception, action, and creative intention within a shared artistic environment.

3.3. Enactive Art Therapy Interfaces and Creative Regulation Systems

Another direction of this research investigates how instrumented creative technologies and co-creative AI systems can support therapeutic creative practice through interactive art-making. Traditional art therapy often relies on qualitative interpretation and therapist observation, making it difficult to systematically study how creative activity contributes to emotional regulation and cognitive change. My work explores instrumented creative interfaces that capture the dynamics of artistic interaction during creative activity, recording behavioral features such as gesture dynamics, drawing tempo, spatial composition, interaction timing, and line density. These interaction traces can be analyzed computationally to model how perception, attention, and engagement evolve during artistic practice. By treating creative activity as a process of regulation within an evolving interaction system, this research aims to develop co-creative therapeutic technologies that support clinicians and artists while providing measurable insight into the dynamics of creative engagement during therapeutic practice.

3.4. Emergence Machine: Enactive AI Decision Architecture

Insights from co-creative systems also motivated the development of the Emergence Machine, an adaptive decision architecture designed to detect structural drift in dynamic environments. The system monitors evolving relationships in streaming data and reorganizes its behavior when underlying structure changes. The architecture has been evaluated across several dynamic datasets, including human-AI co-creative interaction dynamics, physiological signals, large-scale time-series forecasting benchmarks, and financial market data. These experiments examine whether regulatory architectures that monitor structural change—rather than optimizing static predictions—can maintain coherent behavior as environments evolve. The resulting accuracy and error metrics were competitive with state-of-the-art approaches despite requiring no offline training. These results suggest that enactive regulatory architectures may provide a generalizable framework for designing intelligent systems capable of operating in non-stationary environments, including creative interaction systems.

3.5. Unified Enactive AI Architecture

Although these systems operate across domains—including co-creative drawing, dynamic data analysis, and therapeutic creative practice—they are unified by a shared Enactive AI architecture. In this framework, intelligence is modeled as the regulation of agent–environment interaction, enabling systems to adapt their behavior as interaction dynamics evolve. This common foundation allows participatory sense-making and adaptive regulation to be studied across creative, analytical, and therapeutic contexts.

4. Enactive AI Theory and Methods

This research is grounded in theoretical work developed with collaborators in enactive cognitive science and computational creativity. Across several projects, I have proposed frameworks that conceptualize creativity, perception, and cognition as processes that emerge through interaction rather than isolated computation. These contributions include:

Together these frameworks provide a theoretical architecture for designing AI systems capable of sustaining meaningful collaboration with human creators as creative situations evolve. By building computational systems that embody these ideas, theoretical models of creativity and cognition can be tested within real creative environments. Creative technologies thus become experimental laboratories for studying how humans and intelligent systems construct meaning together through interaction.

5. Future Research Directions

Future work expands the Enactive AI research program by investigating how interaction-driven AI systems can support creative practice across increasingly complex environments. These directions extend participatory sense-making from co-creative drawing systems to larger creative media ecosystems, computational models of collaborative creativity, and adaptive interfaces that integrate neural signals with behavioral interaction dynamics.

5.1 Co-Creative AI for Media Production and Immersive Media

One direction investigates how co-creative AI systems can function as collaborators within media production environments such as film, animation, game design, virtual reality, and interactive installations. In these domains, creative work often emerges through iterative exploration rather than linear production pipelines. Enactive AI systems could participate in these processes by suggesting visual variations, narrative possibilities, or stylistic transformations while remaining responsive to human creative intent. In immersive environments, such systems may dynamically respond to participant gestures or artistic input, enabling new forms of participatory storytelling and interactive artistic experiences.

5.2 Modeling Creative Interaction

Another direction expands computational methods for modeling the dynamics of creative collaboration across artistic practices. By analyzing interaction signals—such as drawing gestures, temporal rhythms of creative activity, spatial organization of marks, and patterns of collaborative turn-taking—these systems can capture how creative processes unfold through coordinated interaction between humans and intelligent technologies. Modeling these dynamics enables researchers to study how creative trajectories emerge, how collaborators influence one another’s actions, and how evolving structures within an artwork guide future decisions.

5.3 Enactive Brain–Computer Interfaces for Creative Interaction

A third direction explores the integration of enactive AI models with neural signals from brain–computer interfaces. By combining EEG data with behavioral traces of creative interaction  it may become possible to study how neural activity corresponds to different phases of creative engagement. My prior work on the Emergence Machine suggests that enactive regulatory approaches may be particularly well suited for real-time analysis of EEG signals. Building on this capability, enactive BCI systems could enable adaptive creative interfaces that respond to changes in cognitive workload, attention, and creative exploration as they unfold.

6. Conclusion

The rapid expansion of artificial intelligence across creative industries raises fundamental questions about the future relationship between human creativity and intelligent technologies. Rather than viewing AI as a replacement for human artistic practice, my research explores how intelligent systems can participate meaningfully in creative processes as collaborators. Across my work, I combine theoretical frameworks from enactive cognitive science with experimental creative technology platforms to study and design interaction-driven AI systems that support adaptive sense-making in creative processes. These efforts aim to develop both the conceptual foundations and the technological tools necessary for next-generation creative technologies that expand human creative capacity.

By working across disciplines, I aim to develop co-creative AI systems that expand artistic practice while also deepening our understanding of creativity itself. Ultimately, this research seeks to help shape a future in which artificial intelligence functions not merely as a tool for automated production, but as a creative partner participating with humans in the evolving process of artistic creation.