Myceloom: The Artificial Intelligence of Living Networks

Abstract

Mycelial networks offer profound lessons for AI development transcending biomimicry. Fungal information architectures have executed sophisticated computation and adaptive learning for over 400 million years. This protocol specification articulates principles for AI systems integrating the distributed cognition and computational properties of living networks. The framework synthesizes advances in unconventional computing and reservoir computing. Bio-hybrid robotics and collective intelligence research provide further theoretical grounding. The resulting perspective reveals AI development not as the creation of isolated artificial minds but as the cultivation of symbiotic intelligence networks. These networks honor both computational efficiency and organic wisdom. The specification establishes the AI layer of the Myceloom Protocol (MCP-1) by defining the interface between artificial and biological intelligence within collaborative architectures.

I. Introduction. The Convergence of Biological and Artificial Intelligence

The intelligence landscape of 2025 reveals a distinct shift toward biological substrates. Silicon neural networks dominate contemporary discourse. Fungal networks possess a computational lineage spanning 400 million years.¹ Living mycelium implements Boolean logic circuits and memory formation. Collective decision-making in fungi parallels artificial neural architectures.² The convergence of biological and artificial systems signals a fundamental reconceptualization of information processing.

Current descriptors like "bio-hybrid computing" remain fragmented. Terms such as "fungal electronics" fail to capture the full scope of integration.³ "Myceloom" provides the necessary cohesive framework for systems interfacing with mycelial intelligence.⁴ The concept weaves artificial and biological cognition into a unified approach. The framework draws upon diverse scholarly traditions to articulate principles for designing systems that enhance natural intelligence.

II. The Architecture of Biological Computing

A. Fungal Networks as Computational Substrates

Recent research dissolves the boundary between biological and artificial computation. Andrew Adamatzky and colleagues at the University of the West of England demonstrated that living mycelium networks can implement Boolean circuits through the non-linear transformation of electrical signals.⁵ Fungal composites exhibit "rich dynamics of neuron-like spiking behaviour" and demonstrate genuine computation embedded within living materials.⁶ The work establishes that fungal networks are not merely metaphorically computational but literally implement logical operations.

Mechanisms underlying fungal computation involve calcium waves and electrical potential changes propagating through mycelial networks.⁷ Adamatzky's team showed that these excitation waves can implement computation by encoding Boolean values as spikes of extracellular potential. The researchers "derive sets of two-inputs-on-output logical gates implementable the fungal colony and analyse distributions of the gates."⁸ The distribution of logical functions depends on environmental and physiological conditions of the mycelium-bound composites, providing what the researchers describe as "computational characterisation of the fungal material states."

Fungal networks process information through millions of low-power connections, achieving complex computation without the massive energy requirements of contemporary AI systems.⁹ Artificial neural networks consume gigawatts of power in centralized data centers, whereas mycelium achieves distributed intelligence through inherently efficient parallel processing.¹⁰ The efficiency differential captures the attention of AI researchers seeking alternatives to energy-intensive machine learning architectures.

B. Mycelial Network Topology and Information Processing

Mycelial network topology reveals sophisticated organizational principles paralleling findings in network science. Research on fungal network architecture demonstrates that mycelium develops "scale-free" network properties shared by neural networks, social graphs, and the internet.¹¹ Mark Fricker and Lynne Boddy documented that fungal networks exhibit "efficient foraging strategies" where "adaptive resource allocation" emerges directly from the distributed structure.¹²

The largest known fungal organism, Armillaria bulbosa, spans over 15 hectares and contains an estimated trillion elementary processing units connected through mycelial networks.¹³ This biological computer operates through distributed decision-making, with each hyphal junction acting as a processing node in a vast network architecture predating digital technology.¹⁴ Individual hyphal tips respond to local conditions while contributing to network-wide patterns such as resource allocation, threat response, and growth coordination.

These systems demonstrate "network intelligence," where cognitive capabilities arise from the structure and dynamics of connections rather than individual processing power.¹⁵ Network behavior remains unpredictable from the actions of individual hyphae alone. Intelligence emerges from interactions. The emergent property suggests AI development might focus less on building powerful individual models and more on creating sophisticated networks of interconnected systems demonstrating collective intelligence.

III. Unconventional Computing. Beyond Silicon

A. The Field of Unconventional Computing

The academic field of unconventional computing provides theoretical grounding for Myceloom approaches. Unconventional computing investigates computational processes occurring outside traditional silicon-based architectures, recognizing that computation occurs in biological, chemical, and physical systems.¹⁶ The field establishes that computation constitutes a fundamental property of many dynamical systems rather than belonging exclusively to digital computers.

Research in evolution-in-materio demonstrates that materials contain rich properties exploitable to solve computational problems.¹⁷ This discipline does not abstract computation away from physical substrates but instead exploits intrinsic material dynamics for information processing. Studies show that diverse materials can perform computations, including liquid crystals, nanowire networks, and biological tissues.¹⁸

Biological networks represent potential computational platforms if computation occurs in diverse physical substrates. This recognition motivates research into hybrid systems combining biological and artificial components to exploit each component's unique capabilities.

B. Reservoir Computing. Harnessing Dynamical Systems

Reservoir computing provides a relevant computational framework for understanding Myceloom architectures. Originating in the early 2000s, this approach utilizes dynamical systems as reservoirs where nonlinear generalizations of standard bases adaptively learn spatiotemporal features and hidden patterns in complex time series.¹⁹ RC systems train only a simple readout layer while leaving reservoir dynamics fixed, contrasting sharply with conventional neural networks that require extensive training of all parameters.

A comprehensive Nature Communications review identifies reservoir computing as a promising direction for lightweight AI systems.²⁰ The authors note that biological systems accomplish highly accurate information processing across different scenarios while costing a fraction of the energy needed by large neural networks. Reservoir computing thus offers a computational paradigm aligned with biological efficiency.

Researchers have demonstrated that mycelium can function as a physical reservoir computer, processing information through its natural growth, adaptation patterns, and electrical activity.²¹ Mycelium chips represent hybrid architectures where the biological substrate itself performs information processing, rendering extensive artificial computation unnecessary.²² Electrical activity within the mycelium performs computational transformations on input signals, including voltage spikes and signal propagation. Researchers read out these transformed signals and train a simple linear classifier to accomplish pattern recognition and other computational tasks. The biological substrate performs the heavy computational lifting while the artificial component interprets the results.

IV. Bio-Hybrid Systems. Integration of Living and Artificial

A. Bio-Hybrid Robotics

Bio-hybrid robotics demonstrates practical applications of integrating biological and artificial systems, with a comprehensive npj Robotics review documenting new solutions harnessing the adaptability of living muscles, the sensitivity of living sensory cells, and the computational abilities of living neurons.²³ These systems represent genuine integration rather than mere imitation, allowing biological components to perform functions artificial systems cannot match.

Bio-hybrid robots integrate living organisms—including cells, tissues, and whole organisms—with synthetic materials.²⁴ The resulting systems possess capabilities exceeding either component alone, offering intrinsic softness, environmental safety, and efficient energy conversion superior to traditional robots. Research at ETH Zurich demonstrates bio-hybrid robots capable of various motion abilities, including swimming, bending, rotating, and crawling.²⁵

This approach aligns with Myceloom principles, demonstrating that artificial intelligence need not replace biological intelligence but can be enhanced by it. Bio-hybrid systems maintain living components providing capabilities that remain difficult to achieve artificially, such as adaptation, self-repair, and environmental responsiveness.

B. Fungal Integration in Artificial Systems

Cornell University researchers developed biohybrid robots integrating living mycelium into electronic systems, creating machines that "sense and respond to the environment" through biological computation.²⁶ These systems represent genuine Myceloom architecture where artificial intelligence integrates biological network capabilities rather than simulating them. The living mycelium provides environmental sensing capabilities that inform the robot's behavior.

The integration exploits mycelium's natural responsiveness to environmental stimuli, including light, temperature, chemical gradients, and mechanical pressure.²⁷ Researchers interface artificial systems with biological sensors, creating hybrid architectures where the biological component provides information guiding artificial decision-making.

Machines connected to living mycelial networks exhibit enhanced environmental responsiveness.²⁸ The biological component provides pre-processing, as the mycelium's response to environmental conditions already encodes significant information. The artificial system exploits this data, representing a division of labor where biological and artificial components each perform functions for which they are best suited.

V. Distributed Cognition and Collective Intelligence

A. Theoretical Foundations of Distributed Cognition

The theory of distributed cognition provides conceptual grounding for understanding Myceloom approaches to AI. Developed by Edwin Hutchins and colleagues, this framework proposes that cognitive processes are not confined to individual minds but are distributed across persons, artifacts, and environments.²⁹ Cognition emerges from interactions among components rather than occurring within isolated processors.

Research on collective cognition demonstrates that network topology shapes collective behavior and intelligence.³⁰ A Philosophical Transactions of the Royal Society B review documents how "social network topology shapes collective cognition," where the structure of connections between individuals affects what the collective can accomplish.³¹ This finding has direct implications for AI development, suggesting that system architecture fundamentally constrains collective capabilities.

The distributed cognition framework implies that intelligence need not be concentrated in individual processors but can emerge from appropriately structured networks. This aligns with observations of mycelial networks, where intelligence emerges from distributed interactions in nature rather than centralized control. Developing more intelligent systems requires focusing on network architecture as much as individual component capabilities.

B. AI-Enhanced Collective Intelligence

Recent research examines how AI can participate in collective intelligence systems, with a Patterns review conceptualizing a "multilayer representation of human-AI collective intelligence" comprising cognitive, physical, and information layers.³² The authors propose that humans and AI possess complementary capabilities that, when combined, can surpass the collective intelligence of either in isolation.

This framework aligns with Myceloom principles by positioning AI as a participant in hybrid cognitive systems rather than a replacement for human or biological intelligence. The goal shifts from creating artificial general intelligence that supersedes human cognition to designing AI systems that integrate productively with human, biological, and environmental contexts.

Research on emergent collective memory in decentralized multi-agent AI systems demonstrates how collective capabilities can emerge from distributed interactions.³³ Studies show that multi-agent systems develop collective memory through the interplay between individual agent memory and environmental trace communication, resulting in "spatially distributed collective memory without centralized control." The emergence of collective properties from distributed components mirrors the emergence of network intelligence in mycelial systems.

VI. Neuromorphic Computing. Biological Inspiration for Hardware

A. The Neuromorphic Paradigm

Neuromorphic computing provides a complementary framework for understanding Myceloom approaches, focusing on novel systems that operate at a fraction of the energy of transistor-based computers. These architectures often deviate from the von Neumann model, drawing inspiration from biological principles to create systems that process information more like living organisms.³⁴

Research demonstrates that neuromorphic approaches offers significant advantages for certain computational tasks, with a Springer review noting that reservoir computing on neuromorphic hardware shows particular promise due to "computational efficiency" and the fact that "training amounts to a simple linear regression."³⁵ Both spiking and non-spiking implementations have been developed, demonstrating the versatility of biologically inspired approaches.

The connection to Myceloom thinking involves recognizing that biological systems have already evolved efficient computational strategies. While neuromorphic engineering seeks synthetic substrates better suited to these tasks, biological substrates themselves offer potential solutions.

B. Biocompatible Computing Platforms

Research on reservoir computing with biocompatible organic electrochemical networks demonstrates the practical implementation of biologically inspired computing.³⁶ A Science Advances study reports organic networks demonstrating "complex nonlinear dynamics" and "features typical of biological cortical systems," including recurrency, short-term memory, and E/I balance.³⁷ Researchers achieved 88% accuracy in classifying arrhythmic heartbeats using these platforms.

These systems suggest pathways toward computing platforms that interface directly with biological systems. The authors envision lightweight and noninvasive implants that monitor biosignals and perform 'online' computation without energy-consuming software, representing Myceloom architectures in the literal sense: computation occurring through biocompatible substrates integrating with living systems.³⁸

Development of sustainable memristors from shiitake mycelium demonstrates that fungal materials can provide scalable platforms for neuromorphic tasks.³⁹ Biological processors offer alternatives to resource-intensive silicon architectures, demonstrating computational capabilities paralleling artificial neural networks. The material basis of computation thus shifts from extracted minerals to cultivated biological materials.

VII. The Myceloom Framework. Principles for Symbiotic Intelligence

A. Core Principles

The Myceloom framework articulates principles for AI development that integrate biological and artificial intelligence, emphasizing distributed processing, biological integration, emergent intelligence, and symbiotic enhancement.

Distributed Processing: Myceloom architectures prioritize distributed rather than centralized computation, where intelligence emerges from interactions among network components. This approach offers resilience, allowing the system to continue functioning despite individual component failures, and efficiency, as computation occurs locally where information resides.

Biological Integration: These systems maintain living components performing functions for which they correspond uniquely, exploiting capabilities difficult to achieve artificially. Examples include adaptation, self-repair, environmental responsiveness, and energy efficiency.

Emergent Intelligence: Myceloom systems operate on the premise that collective intelligence can exceed individual component capabilities. The goal lies in optimizing network architecture for emergent collective capabilities rather than maximized individual processor power, aligning with both mycelial network organization and distributed cognition theory.

Symbiotic Enhancement: Myceloom AI systems aim to enhance rather than replace other forms of intelligence, seeking productive integration with human cognition, biological systems, and environmental contexts.

B. Computational Symbiosis: The Living Machine Interface

Myceloom systems recognize intelligence as inherently collaborative rather than competitive. Recent advances in fungal computing reveal that living mycelium can implement reservoir computing where the biological substrate itself performs information processing.⁴⁰ These hybrid architectures represent a fundamental shift: AI becomes an enhancement of biological capabilities rather than a replacement.

The symbiotic approach offers solutions to pressing AI challenges. While contemporary machine learning systems require massive datasets and energy-intensive training processes, Myceloom architectures exploit biological intelligence "trained" by evolutionary optimization.⁴¹ Research shows that AI systems inspired by mycelial principles demonstrate superior resilience, energy efficiency, and adaptive capacity compared to traditional centralized AI systems.⁴²

This symbiosis functions bidirectionally: artificial augmentation enhances biological systems just as AI systems benefit from biological integration. Bio-hybrid robots demonstrate enhanced capabilities through combining biological and artificial components, producing systems with capabilities exceeding either component alone—a genuine symbiosis rather than parasitism or competition.⁴³

VIII. Applications and Implications

A. Practical Implementations

Current research demonstrates practical applications for Myceloom architectures across multiple domains. Fungal computing researchers show that mycelial networks can solve complex optimization problems, including shortest path calculations, network topology optimization, and adaptive resource allocation.⁴⁴ These biological computers operate through environmental programming, where modifying growth conditions effectively reprograms network geometry and computational behavior.

Bio-hybrid robotics research demonstrates applications in environmental monitoring, drug delivery, and search-and-rescue operations.⁴⁵ Bio-hybrid systems offer distinct advantages in wet, constrained, or biologically sensitive contexts where living components provide capabilities difficult to achieve artificially, such as self-repair, environmental responsiveness, and efficient energy utilization.

Machine learning researchers implement Myceloom principles through decentralized AI architectures that distribute computational tasks across networks of simple processors, mimicking how fungi allocate resources based on environmental needs.⁴⁶ These systems demonstrate the democratizing potential of Myceloom architecture, offering sophisticated computational capabilities with decreased infrastructure requirements.

B. Sustainability and Energy Efficiency

Biological computing offers significant sustainability advantages over contemporary large language models, which consume substantial energy during training and inference.⁴⁷ As the environmental impact of AI development becomes a growing concern, Myceloom approaches offer an alternative trajectory with a reduced ecological footprint.

Biological computing systems operate at significantly lower energy levels than silicon alternatives, processing information through chemical and electrical signaling that requires minimal energy input.⁴⁸ While biological systems cannot match silicon's raw computational speed, they achieve remarkable efficiency in sensing, adaptation, and pattern recognition tasks where efficiency takes precedence over raw processing power.

Cultivating computational materials rather than extracting them represents a paradigm shift in how computing infrastructure relates to environmental systems. Myceloom approaches suggest growing computational substrates from renewable biological materials, aligning computing development with broader sustainability goals.⁴⁹

IX. Challenges and Future Directions

A. Technical Challenges

Significant technical challenges remain in developing Myceloom architectures, particularly the "timescale problem" where matching computational timescales between biological and artificial systems proves complex.⁵⁰ Biological processes often operate at different temporal scales than electronic computation, requiring careful interface design.

Maintaining living systems within artificial frameworks also requires addressing biological needs such as nutrition, environmental conditions, and waste removal.⁵¹ Unlike silicon components that remain stable indefinitely, biological components have lifespans and require ongoing maintenance. Developing sturdy bio-hybrid systems thus requires solving fundamental problems of biological sustainability.

Scalability presents another significant hurdle. While laboratory demonstrations prove biological computation principles, scaling to practical applications requires addressing biological variability, environmental sensitivity, and integration complexity.⁵² The transition from proof-of-concept to deployable systems represents a major engineering challenge.

B. Ethical Considerations

Integrating living systems into computing architectures raises ethical questions regarding the moral status of living components and the appropriateness of creating novel bio-artificial hybrids.⁵³ Using fungal networks involves different considerations than using animal tissues or neural cells, necessitating ethical frameworks that account for diverse biological components.⁵⁴

Environmental implications extend beyond sustainability benefits to include potential risks, such as unintended ecological consequences from releasing bio-hybrid systems. Responsible development requires careful consideration of containment, environmental impact assessment, and end-of-life management for biological computing components.⁵⁵

X. Conclusion. The Philosophy of Symbiotic Intelligence

"Myceloom" provides essential terminology for navigating the convergence of biological and artificial intelligence, conveying qualities that are biological, collaborative, adaptive, and intelligent. This precision enables clearer thinking about AI development that honors both computational efficiency and organic wisdom.

The convergence of fungal computing research, reservoir computing theory, bio-hybrid robotics, and collective intelligence scholarship points toward a unified understanding: intelligence appears inherently distributed and potentially symbiotic. Individual processors achieve their fullest expression through integration into collaborative networks, which in turn achieve their highest capabilities through the appropriate composition and connection of diverse components. The health and capability of each enables the health and capability of all.

As researchers advance toward sophisticated artificial intelligence, mycelial networks beneath forest floors offer profound lessons about distributed cognition, adaptive learning, and symbiotic collaboration. The future of AI may lie not in perfecting isolated artificial minds but in weaving them into living networks connecting all intelligent life. The Myceloom framework captures this evolution: artificial intelligence systems growing like fungi, adapting like living networks, and demonstrating the collaborative intelligence necessary for addressing complex global challenges.

The convergence of ancient biological wisdom and cutting-edge technology offers a pathway where AI enhances rather than replaces the natural intelligence of living systems. AI developed according to Myceloom principles can achieve heights that neither isolated artificial systems nor unaugmented biological systems can accomplish alone. This vision represents not a utopian aspiration but an empirically grounded possibility documented by laboratory demonstrations, theoretical frameworks, and emerging practical applications. The future substrate for artificial intelligence lies not in competition with biological intelligence but in connection; not in replacement but in enhancement; not in isolation but in integration.

Works Cited

Adamatzky, Andrew. "Towards Fungal Computer." Interface Focus 8, no. 6 (2018): 20180029.

Adamatzky, Andrew, ed. Advances in Unconventional Computing. Vol. 22, Emergence, Complexity and Computation. Cham: Springer, 2017.

Adamatzky, Andrew, Phil Ayres, Alexander E. Beasley, Nic Roberts, and Han A. B. Wösten. "Logics in Fungal Mycelium Networks." Logica Universalis 16, no. 4 (2022): 655–669.

Adamatzky, Andrew, Martin Tegelaar, Han A. B. Wösten, Anna L. Powell, Alexander E. Beasley, and Richard Mayne. "On Boolean Gates in Fungal Colony." Biosystems 193 (2020): 104138.

Boddy, Lynne. "Fungal Behavior: A New Front in Behavioral Ecology." Fungal Ecology 32 (2018): 92–101.

Coman, Alin, Ida Momennejad, Rae D. Drach, and Andra Geana. "Collective Minds: Social Network Topology Shapes Collective Cognition." Philosophical Transactions of the Royal Society B 377 (2022): 20200315.

Cucchi, Matteo, Steven Abreu, Giuseppe Ciccone, Daniel Brunner, and Hans Kleemann. "Reservoir Computing with Biocompatible Organic Electrochemical Networks for Brain-Inspired Biosignal Classification." Science Advances 7, no. 34 (2021): eabh0693.

Cucchi, Matteo, Steven Abreu, Giuseppe Ciccone, Daniel Brunner, and Hans Kleemann. "Reservoir Computing with Biocompatible Organic Electrochemical Networks for Brain-Inspired Biosignal Classification." Science Advances 7, no. 34 (2021): eabh0693.

Cui, Hao, Tao Zhou, Yidong Chai, Haifeng Ling, Hai-Tao Zhang, and Jeffrey V. Nickerson. "AI-Enhanced Collective Intelligence." Patterns 5, no. 9 (2024): 101074.

Dale, Matthew, Julian F. Miller, and Susan Stepney. "Reservoir Computing as a Model for In-Materio Computing." In Advances in Unconventional Computing, vol. 22, edited by Andrew Adamatzky, 533–571. Cham: Springer, 2017.

Filippi, Matteo, Thomas Buchner, Oncay Yasa, Stefan Weirich, and Robert K. Katzschmann. "Microfluidic Tissue Engineering and Bio-actuation." Advanced Materials 34 (2022): 2108427.

Fricker, Mark D., Luke L. M. Heaton, Nick S. Jones, and Lynne Boddy. "The Mycelium as a Network." In The Fungal Kingdom, edited by Joseph Heitman et al., 335–367. Washington, DC: ASM Press, 2017.

Hadaeghi, Fatemeh. "Neuromorphic Electronic Systems for Reservoir Computing." In Reservoir Computing: Theory, Physical Implementations, and Applications, edited by Kohei Nakajima and Ingo Fischer, 221–249. Singapore: Springer, 2021.

Harding, Simon, and Julian F. Miller. "Evolution-in-Materio: Evolving Computation in Materials." Evolutionary Intelligence 1 (2008): 85–98.

Hollan, James D., Edwin Hutchins, and David Kirsch. "Distributed Cognition: Toward a New Foundation for Human-Computer Interaction Research." ACM Transactions on Computer-Human Interaction 7, no. 2 (2000): 174–196.

Hutchins, Edwin. Cognition in the Wild. Cambridge, MA: MIT Press, 1995.

Khushiyant. "Emergent Collective Memory in Decentralized Multi-Agent AI Systems." arXiv preprint (2025).

LaRocco, John, et al. "Sustainable Memristors from Shiitake Mycelium for High-Frequency Neuromorphic Computing." bioRxiv (2025).

Mestre, Rafael, Tania Patiño, and Samuel Sánchez. "Ethics and Responsibility in Biohybrid Robotics Research." Proceedings of the National Academy of Sciences 121, no. 31 (2024): e2310458121.

Mishra, Anand K., et al. "Sensorimotor Control of Robots Mediated by Electrophysiological Measurements of Fungal Mycelia." Science Robotics 9, no. 93 (2024): eadk8019.

Money, Nicholas P. "Hyphal and Mycelial Consciousness: The Concept of the Fungal Mind." Fungal Biology 125, no. 4 (2021): 257–259.

Roberts, Nic, and Andrew Adamatzky. "Mining Logical Circuits in Fungi." Scientific Reports 12 (2022): 15389.

Sinha, Shaurya, et al. "Biohybrid Living Robotics: A Comprehensive Review of Recent Advances, Technological Innovation, and Future Prospects." npj Robotics 3 (2025): 12.

Smith, Myra L., Johann N. Bruhn, and James B. Anderson. "The Fungus Armillaria Bulbosa Is Among the Largest and Oldest Living Organisms." Nature 356, no. 6368 (1992): 428–431.

Strubell, Emma, Ananya Ganesh, and Andrew McCallum. "Energy and Policy Considerations for Deep Learning in NLP." Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (2019): 3645–3650.

Webster-Wood, Victoria A., et al. "Biohybrid Robots: Recent Progress, Challenges, and Perspectives." Bioinspiration & Biomimetics 18, no. 1 (2023): 015001.

Yan, Min, et al. "Emerging Opportunities and Challenges for the Future of Reservoir Computing." Nature Communications 15 (2024): 2056.

Footnotes