Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 110 of the 110 references it cites were found in a registry or an index; 43 of the 125 sources its claims cite were checked for support.

Wetware network-based AI: a chemical approach to embodied cognition for robotics and artificial intelligence

Open access
Damiano, Luisa, Fleres, Antonio, Roli, Andrea, Stano, Pasquale · 2026 · Frontiers in Robotics and AI

Abstract

Wetware Network-Based Artificial Intelligence (WNAI) introduces a new approach to robotic cognition and artificial intelligence: autonomous cognitive agents built from synthetic chemical networks. Rooted in Wetware Neuromorphic Engineering , WNAI shifts the focus of this emerging field from disembodied computation and biological mimicry to reticular chemical self-organization as a substrate for cognition. At the theoretical level , WNAI integrates insights from network cybernetics, autopoietic theory and enaction to frame cognition as a materially grounded, emergent phenomenon. At the heuristic level , WNAI defines its role as complementary to existing leading approaches. On the one hand, it complements embodied AI and xenobotics by expanding the design space of artificial embodied cognition into fully synthetic domains. On the other hand, it engages in mutual exchange with neural network architectures, advancing cross-substrate principles of network-based cognition. At the technological level , WNAI offers a roadmap for implementing chemical neural networks and protocellular agents, with potential applications in robotic systems requiring minimal, adaptive, and substrate-sensitive intelligence. By situating wetware neuromorphic engineering within the broader landscape of robotics and AI, this article outlines a programmatic framework that highlights its potential to expand artificial cognition beyond silicon and biohybrid systems.

Overview

CoChat Summary

TL;DR

Wetware Network-Based Artificial Intelligence (WNAI) is a programmatic framework that proposes synthetic chemical networks as a substrate for embodied cognition, complementing embodied AI and xenobotic AI. It defines operational criteria for minimal sense-making in chemical systems and outlines a roadmap toward chemical neural networks and protocellular agents for robotics and AI.

Key findings

  • WNAI is positioned as a third complementary strategy alongside electromechanical embodied AI and xenobotic AI, using non-living synthetic chemical networks with bottom-up self-organization rather than engineered modularity or biological tissues (Table 1).
  • The paper operationally defines sense-making as a chemical network’s emergent capacity to modulate its own structure through self-regulation under perturbation while maintaining organization and identity; it lists observables such as structural persistence, adaptive modulation of couplings, energy-gradient maintenance, boundary plasticity, and emergent pattern differentiation (Table 2).
  • It proposes five guiding design concepts—organization, emergence, autonomy, co-evolution, and closure—and a tentative roadmap with four directions: protocellular construction, emergent wetware dynamics, situated system–environment interaction, and observables/metrics; Table 3 lists five strategic axes.
  • It illustrates a minimal WNAI implementation with Luisi and Varela’s autopoietic reverse micelles: 50 mM sodium octanoate, aqueous LiOH, 9:1 (v/v) isooctane/1-octanol, reverse micelles of ca. 5 nm diameter, with octyl octanoate saponification driving growth and division. Proposed applications include microrobotics for precision medicine, soft robotics in extreme environments, and ecological monitoring/remediation.

Methods

  • Design: Hypothesis and Theory article; a conceptual and programmatic framework, not an empirical study.
  • Data or sample: no experimental sample or dataset; the paper synthesizes literature from network cybernetics, autopoiesis/enaction, synthetic biology, and systems chemistry, and cites prior experimental models such as reverse micelles.
  • Main analysis: conceptual argument, comparative tables (Tables 1–3), operational criteria for minimal cognition, and a tentative roadmap; no statistical or experimental analysis is reported.

Limitations

  • The paper states technological limitations related to stability at molecular and supramolecular levels, timescales of self-regulation versus perturbation, noise, and scalability; the auxiliary-network strategy is described as a pragmatic compromise that may constrain core network dynamics and plasticity.
  • The paper states that application scenarios are “largely speculative” and represent potential outcomes of later research; full autopoietic chemical AI remains practically challenging, so initial work targets autonomous rather than autopoietic systems.
  • Evident from the text: no original experimental validation or quantitative results are reported, and the proposed operational criteria and roadmap remain to be implemented.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3389/frobt.2025.1694338

Cites

110 of 110 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

FIGURE 1 (A) Schematic drawing of two networks (the “system” and the “environment” in reciprocal interaction. Inspired by (Ashby, 1960), p. 70. (B) To facilitate experimental approaches, some necessary, yet not dynamics-determining processes, can be offloaded to an auxiliary network, which constitutes a sort of inner reservoir for pseudo-stationary exchanges. The auxiliary network should not contribute, or contribute only indirectly, to the autonomous and self-regulating mechanisms of interest.
FIGURE 1 (A) Schematic drawing of two networks (the “system” and the “environment” in reciprocal interaction. Inspired by (Ashby, 1960), p. 70. (B) To facilitate experimental approaches, some necessary, yet not dynamics-determining processes, can be offloaded to an auxiliary network, which constitutes a sort of inner reservoir for pseudo-stationary exchanges. The auxiliary network should not contribute, or contribute only indirectly, to the autonomous and self-regulating mechanisms of interest.CC BY · PubMed Central
FIGURE 2 An illustrative version of the general scheme shown in Figure 1A. The network represents a possible set of interconnected variables and processes that can be identified in the autopoietic reverse micelles (Luisi and Varela, 1989; Bachmann et al., 1990), which can serve as a minimal example of WNAI implementations. Variables and processes on the left of the dashed line refer to the environment, while the others refer to the system. Events occurring at the molecular and supramolecular domains are intertwined: the propensities and the rates of processes are per se not constant, but depends on the value of certain variables, which in turn depends on the ongoing processes, possibly on a different scale. The composition of the system, e.g., of the reverse micelle amphiphilic boundary, determines the exact value of free energy, ultimately affecting physical as well as chemical processes occurring on the boundary region. Gradients of intensive variables such as chemical potential or surface tension and their conjugated flows are also related. A detailed description of how the system works, in summary, is given in the text (Section 5.4). In this network, M: micelle; P: precursor (octyl octanoate); S: micelle-forming amphiphilic compound (octanoate); Q: byproduct of the P-to-S transformation (1-octanol), also participating to the boundary and present in outer environment, together with isooctane; LiOH: lithium hydroxide (the inner reactant); w 0: [water]/[S] ratio, a major determinant of micellar structure/stability (but it is also context-dependent, i.e., micelles formed in two different solvents, even characterized by the same w 0, are not necessarily similar). It should also be noted that the extremely small size of these systems is expected to introduce stochastic behavior across all processes (e.g., “redistribution of solutes”, “redistribution of boundary components”). The network has been drawn with illustrative purposes, and it is not supposed to be exhaustive of all possible causal connections (e.g., sodium/lithium competition as counterions has not been included). Another illustrative reaction network, based on a simplified model of protein-synthesizing artificial cells, can be found in Gentili and Stano (2023a).
FIGURE 2 An illustrative version of the general scheme shown in Figure 1A. The network represents a possible set of interconnected variables and processes that can be identified in the autopoietic reverse micelles (Luisi and Varela, 1989; Bachmann et al., 1990), which can serve as a minimal example of WNAI implementations. Variables and processes on the left of the dashed line refer to the environment, while the others refer to the system. Events occurring at the molecular and supramolecular domains are intertwined: the propensities and the rates of processes are per se not constant, but depends on the value of certain variables, which in turn depends on the ongoing processes, possibly on a different scale. The composition of the system, e.g., of the reverse micelle amphiphilic boundary, determines the exact value of free energy, ultimately affecting physical as well as chemical processes occurring on the boundary region. Gradients of intensive variables such as chemical potential or surface tension and their conjugated flows are also related. A detailed description of how the system works, in summary, is given in the text (Section 5.4). In this network, M: micelle; P: precursor (octyl octanoate); S: micelle-forming amphiphilic compound (octanoate); Q: byproduct of the P-to-S transformation (1-octanol), also participating to the boundary and present in outer environment, together with isooctane; LiOH: lithium hydroxide (the inner reactant); w 0: [water]/[S] ratio, a major determinant of micellar structure/stability (but it is also context-dependent, i.e., micelles formed in two different solvents, even characterized by the same w 0, are not necessarily similar). It should also be noted that the extremely small size of these systems is expected to introduce stochastic behavior across all processes (e.g., “redistribution of solutes”, “redistribution of boundary components”). The network has been drawn with illustrative purposes, and it is not supposed to be exhaustive of all possible causal connections (e.g., sodium/lithium competition as counterions has not been included). Another illustrative reaction network, based on a simplified model of protein-synthesizing artificial cells, can be found in Gentili and Stano (2023a).CC BY · PubMed Central

Verification

Claims the paper makes

29 Valid · 14 Not valid · 34 Not checked · 100 Own results and interpretations

62 more claims · 14 with their cited sources checked

Create a free account to see all 177 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 91
  • It may state two separate claims as one. 40

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Citations

Cites

Show all 106 references

Research works printed but not linked: 11

  1. Bray, 2009, Wetware. A computer in every living cells
  2. Guidoni, 2018, Verso il Robot sapiens, Monte Compatri (Roma) Ed. Control
  3. Kauffman, 2016, Humanity in a creative universe
  4. LeCun, 2025, Speech at “debating technology” session of the 2025 annual meeting of the world economic forum
  5. Maturana, 1973, De Máquinas y Seres Vivos: Una Teoría sobre la Organización Biológica
  6. McCulloch, 1965, Embodiments of mind
  7. Morowitz, 1979, Energy flow in biology
  8. Piaget, 1967, Biologie et connaissance: Essai sur les relations entre les régulations organiques et les processus cognitifs
  9. Varela, 1979, Principles of biological autonomy
  10. Varela, 1995, The emergent self, 209
  11. von Foerster, 1960, On self-organizing systems and their environments, 31

Cited by

See all 2

Metadata

Authors
Damiano, L., Fleres, A., Roli, A., & Stano, P.
Affiliations
  • IULM University (IT)
  • European Centre for Living Technology (IT)
  • University of Salento (IT)
Corresponding author
Damiano, Luisa
Venue
Frontiers in Robotics and AI
Published
2026-01-05
Volume
12
Pages
1694338-1694338
ISSN
2296-9144 (ISSN-L)
OpenAlex
W7118171687
Cited by
2
References
106
Citations by year
2026: 2
FWCI
2.97
Type
journal article
Language
en
Open access
Locations
  • Frontiers in Robotics and AI (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • Archivio istituzionale della ricerca (Alma Mater Studiorum Università di Bologna) (submittedVersion, cc-by)
  • AIRUS (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Ministero dell’Istruzione, dell’Università e della Ricerca
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Psychiatry, Mental Health, Neuroscience
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Embodied and Extended Cognition
  • Social Sciences › Arts and Humanities › History and Philosophy of Science › Cybernetics and Technology in Society
Keywords
embodied cognition, protocells, autopoiesis, enaction, emergent cognition, artificial cognition

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Wetware network-based AI: a chemical approach to embodied cognition for robotics and artificial intelligence. https://app.cochat.ai/papers/10.3389/frobt.2025.1694338
The paper
Damiano, L., Fleres, A., Roli, A., & Stano, P. (2026). Wetware network-based AI: a chemical approach to embodied cognition for robotics and artificial intelligence. Frontiers in Robotics and AI. https://doi.org/10.3389/frobt.2025.1694338

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.