Rhizomic Mind: Thinking Outside the Hierarchy

A reflection on the philosophy of Deleuze & Guattari and its surprising resonances with my AI Governance framework

AI Image by Author - Chat GPT

Dear readers,

In the continuous unfolding of my conceptual work, a path that defines and redefines itself with each reflection, a surprising and stimulating consonance has recently emerged. As I deepened my research to enhance the approach to artificial intelligence governance that I am developing—particularly by exploring the dynamics of lateral thinking and the applications of chaos theory to navigate complexity—I found myself rediscovering more profoundly the concept of the "rhizome," as brilliantly articulated by the philosophers Gilles Deleuze and Félix Guattari. The encounter was striking: I perceived a powerful echo, a deep resonance with the very structure and aspirations of my Custos AI architecture. However, such a philosophical figure had not consciously guided my initial design.

The rhizome, for those not deeply familiar with it, is proposed by Deleuze and Guattari as a model of thought radically alternative to traditional linear, hierarchical logics—those they define as "arborescent." If the tree evokes a central trunk, ordered roots, and predictable branches, the rhizome, conversely, expands horizontally, like a subterranean and multiform network, devoid of a centre or a unique command point. Each of its elements can dynamically connect to any other, often in unexpected and non-sequential ways, weaving a complex web of paths and interactions. The rhizome is not a unit that divides into subordinate parts, but a multiplicity in constant becoming, always "in the middle," an open system capable of connecting heterogeneous elements. Deleuze and Guattari describe it not as a "tracing," which passively reproduces pre-existing structures, but as a "map": a map always modifiable, orientable in all directions, which describes a landscape of possibilities and constantly encourages the creation of "lines of flight" towards unexplored territories and new configurations of meaning.

It was, then, an after-the-fact discovery, a true 'aha moment'. Precisely as I was working on how chaos theory could inform the capacity of my endeavor to generate significant "drifts" from an ethical query, and how lateral thinking could support the mapping of a vast horizon of possibilities rather than convergence towards a single, predetermined answer, I grasped the profound affinity with rhizomatic principles. The proposal, in its attempt to govern the intrinsic complexity of advanced AI systems, seems inherently to escape a purely arborescent and hierarchical logic. Its aspiration is not to impose order from above, but to trace and navigate a network of ethical implications, acentric and multi-connected, where each "Ethical Challenge" can open up to multiple interdependent lines of inquiry, much like a rhizome.

Let us consider, for example, the component I have named "Preventive Chaotic Intelligence": it is specifically conceived to explore a myriad of simulated scenarios, potential deviations, veritable "lines of flight" that an AI system, given a certain initial condition, might undertake. The objective is not to identify a single "root cause" of a problem, but rather to map a "plateau"—to use Deleuze and Guattari's term—of risks, vulnerabilities, and systemic implications. A plateau, in their thinking, represents a zone of continuous intensity, a field of forces in constant transformation, without a fixed peak or definitive resolution. Consequently, ethical analysis within the Custos AI ecosystem does not so much aim to provide a "final solution" that is static, but rather to promote a dynamic and adaptive understanding of an ever-evolving risk landscape.

The emphasis on "Shepherd Algorithms" and the meticulous collection of "Ethical Decision Crash Recordings" (EDCR) within the Ethical Archives, both National (AERN) and Central (CEA), can also be interpreted through such a rhizomatic lens. This is not a mere accumulation of errors or isolated incidents, but the construction of a collective, distributed, and interconnected memory: a dynamic "map" of ethical drifts observed in the real world. A similar constant mapping, fed by a multiplicity of sources, is designed to favour the emergence of patterns, correlations, and systemic understandings that a fragmented or purely hierarchical vision would hardly grasp. The system learns and adapts just as a rhizome extends, assimilates, and creates new connections in response to its environment.

This unexpected and profound convergence, stemming from my immersion in lateral thinking and the applications of chaos theory to governance, suggests to me that, perhaps, to address the elusive, interdependent, and ever-evolving nature of advanced artificial intelligence, the most fruitful and resilient oversight approaches can only be "rhizomatic" themselves. We need mechanisms capable of embracing multiplicity, fostering transversal connections between different domains of knowledge, operating without an omnipotent centralised control, and possessing a rooted capacity for continuous adaptation and exploration.

Such an insight into rhizomatic resonance is proving to be a particularly fertile line of thought. I am currently exploring ways to give these dynamics a more concrete and tangible form, aiming to translate the complexity of chaos exploration and the inherent possibilities of my Custos AI scheme into something new, perhaps even a "ludic complement" dedicated to it. An itinerary that, I hope, may offer fresh perspectives and direct modes of experimentation with these fundamental ideas for the future of AI.

---

Discover more stories and reflections in my books, or connect with my professional journey on LinkedIn. I always value your input; please feel free to reach out with feedback, suggestions, or inquiries to cosmicdancerpodcast@gmail.com.