Agentic Digital Twins: A Taxonomy of Capabilities for Understanding Possible Futures

TL;DR

The paper addresses how digital twins can become agentic, extending beyond representation to co-constitute the systems they model. It introduces a three-dimensional taxonomy—locus of agency, tightness of coupling, and model evolution—yielding $27$ configurations and nine illustrative cases organized into The Present, The Threshold, and The Frontier. Using traffic navigation and performative prediction theory, it traces a progression from passive tools to reconstructive ontologies and discusses risks such as performative lock-in and epistemic inscrutability, along with governance strategies. The work provides a structured framework to anticipate, evaluate, and guide the development of agentic DTs, emphasizing anticipatory governance, multi-stakeholder oversight, and safeguards to shape beneficial futures while preserving human agency.

Abstract

As digital twins (DTs) evolve to become more agentic through the integration of artificial intelligence (AI), they acquire capabilities that extend beyond dynamic representation of their target systems. This paper presents a taxonomy of agentic DTs organised around three fundamental dimensions: the locus of agency (external, internal, distributed), the tightness of coupling (loose, tight, constitutive), and model evolution (static, adaptive, reconstructive). From the resulting 27-configuration space, we identify nine illustrative configurations grouped into three clusters: "The Present" (existing tools and emerging steering systems), "The Threshold" (where emergent properties appear and coupling becomes constitutive), and "The Frontier" (where systems gain reconstructive capabilities). Our analysis explores how agentic DTs exercise performative power--not merely representing physical systems but actively participating in constituting them. Using traffic navigation systems as examples, we show how even passive tools can exhibit emergent performativity, while advanced configurations risk performative lock-in. Drawing on performative prediction theory, we trace a progression from passive tools through active steering to ontological reconstruction, examining how constitutive coupling enables systems to create self-validating realities. Understanding these configurations is essential for navigating the transformation from DTs as mirror worlds to DTs as architects of new ontologies.

Figures (6)

• Figure 1: The three levels of agency: external agency where agents control digital twins as tools, internal agency where embedded agents make autonomous decisions, and distributed agency where control emerges from networked interactions.
• Figure 2: The three levels of coupling: loose coupling with periodic updates and sampled data, tight coupling with real-time control and continuous feedback, and constitutive coupling where digital twin and physical system constitute a new higher-level hybrid system with emergent properties.
• Figure 3: The three levels of model evolution: static models with fixed parameters, adaptive models that optimise predictions through feedback learning, and reconstructive models with evolving ontologies that reconstitute the system being modelled.
• Figure 4: Three-dimensional coordinate system showing three example DT configurations
• Figure 5: A six-step progression path through the conceptual space from Tool Configuration (1) to Reconstructive Assemblage (9). Transition labels show driving factors using deliberative notation. The colour gradient from green to red indicates progression toward configurations with greater autonomy and reduced reversibility.