Enactive AI Research Program

Modern intelligent systems increasingly operate in environments characterized by structural change rather than stable equilibrium. Financial markets, organizational workflows, and distributed human–AI systems evolve continuously, often invalidating static assumptions embedded in traditional optimization pipelines. As correlations shift and contextual structures reorganize, system performance depends less on isolated predictive accuracy and more on the ability to adapt behavior under changing conditions.

This dossier introduces an architectural perspective grounded in the cognitive science paradigm of enaction. From this perspective, intelligence arises through ongoing interaction between an agent and its environment. Effective systems therefore maintain coherence not by optimizing fixed objectives, but by continuously interpreting changes in their surroundings and reorganizing their engagement accordingly.

The work presented here operationalizes this perspective by treating structural drift as an informational signal rather than noise. When environmental structure shifts, adaptive systems must update their interpretation of the situation and adjust their behavioral posture. This process is supported through regulatory feedback loops that couple perception, interpretation, and action.

This research program explores how:

To investigate these ideas empirically, I design experimental computational platforms that operate in real-time environments. One such system, the Emergence Machine, explores how adaptive decision architectures can detect structural drift in streaming data (e.g. financial markets, physiological signals, energy infrastructure) and reorganize behavioral engagement without relying on any offline training.

Rather than positioning AI as an isolated optimization engine, this work frames intelligent systems as participants within distributed processes of shared sense-making and adaptive coordination. The objective is not automation alone, but the design of systems capable of sustaining coherent behavior in dynamic environments.

This research explores how intelligent systems—both human and artificial—maintain coherent behavior under conditions of structural change. Rather than treating intelligence as static optimization over fixed assumptions, my work develops architectures that regulate behavior through continuous interaction with evolving environments. This perspective draws on cognitive science frameworks such as enaction, distributed cognition, and dynamical systems theory, translating these ideas into computational architectures for adaptive decision systems.

Across this body of work, Davis et al. designs experimental platforms that allow these theoretical principles to be studied empirically. These systems operate and learn in real time, interpreting structural change in their environment and reorganizing their engagement accordingly. One such platform, the Emergence Machine, explores how adaptive systems can detect structural drift in streaming data (e.g. EEG, EKG, energy/infrastructure signals, stocks and cryptocurrency) and regulate behavioral posture without reliance on fixed offline training assumptions (i.e. no offline training). This work contributes to a broader research program focused on adaptive AI architectures capable of sustaining effective behavior in non-stationary environments.


Core Research Contributions

1. Cognitive Architecture Design

My work develops architectural principles for adaptive intelligent systems informed by enactive and dynamical approaches to cognition.

Key contributions include:

2. Human–AI Interactive Systems

I design and empirically study interactive systems that enable humans and AI to coordinate effectively in real-time environments.

This research includes:

3. Experimental Platforms for Adaptive AI

My research is supported by the development of experimental computational platforms that explore next-generation adaptive architectures.

These platforms allow investigation of:

Through these experimental systems, theoretical principles from cognitive science can be evaluated in operational environments.

Key Innovations in Co-Creative AI and Enactive AI

My work advances a human-centered paradigm for adaptive intelligent systems grounded in enaction — the view that cognition arises through ongoing interaction between agent and environment. Rather than treating intelligence as static representation or offline optimization, enactive systems build meaning through continuous engagement, regulation, and sense-making.

Building on this foundation, I have developed several interrelated cognitive frameworks:

These theories are not purely conceptual. They inform the design of computational systems that operationalize enactive regulation.

The Emergence Machine represents one such implementation: an experimental Enactive AI system that operates entirely online, detecting shifts in structural conditions and adjusting behavioral posture accordingly. Rather than optimizing a fixed objective, it models context, recognizes regime change, and adapts its engagement over time. The goal is not improved prediction alone, but sustained effectiveness under dynamic conditions.

In parallel, my work in co-creative AI positions human–AI interaction not as automation, but as participatory coordination. Through the design and empirical study of real-time co-creative drawing systems, I have developed methods to quantify interaction rhythms, adaptive turn-taking, and sense-making trajectories between human and machine. This paradigm reframes AI as a collaborative partner embedded within evolving cognitive systems.

Across these projects, the unifying aim is to design intelligent systems that regulate, interpret, and coordinate within changing environments — supporting human judgment rather than replacing it. This approach contributes to the development of next-generation adaptive AI architectures capable of operating coherently across domains characterized by uncertainty, drift, and structural change.