The vision is not a vague ecological metaphor. Armstrong and Moodley ground their framework in technologies that have already been validated in laboratories and pilot projects, including bioelectrochemical systems such as microbial fuel cells, which generate electricity directly from the metabolic activity of bacteria breaking down organic waste. The idea of architecture that behaves like a living organism has deep roots, from the Japanese Metabolist movement of the 1960s to Gordon Pask’s electrochemical machines and the cybernetic experiments of Nicholas Negroponte. What is new in the CPLS proposal is the synthesis of that biological ambition with modern artificial intelligence, and specifically with the semantic technologies that allow machines to reason over knowledge rather than simply react to sensor data.

At the heart of the framework is a tripartite symbiosis, an ecosystem of three classes of actors. The first is a metabolic ‘gut’: integrated bioreactors embedded in the fabric of the building that digest waste streams, transform them into energy and useful materials, and give the house a functioning physiology. The second is a cognitive ‘brain’: a cognitive digital twin, a computational model of the home that does not merely mirror its physical state but can explain its own reasoning, learn from experience, and coordinate the building’s biological and technical subsystems. The third, and crucially not the least, is the human occupant, repositioned from a passive consumer of automated services into a literate steward who participates in and understands the home’s living processes.

The technical linchpin that makes this symbiosis computable is a proposed hybrid ontology. In computer science, an ontology is a formal, machine-readable specification of the concepts in a domain and the relationships between them. Existing building ontologies, such as the W3C’s BOT building topology ontology, the SSN/SOSA sensor ontologies and the ETSI SAREF family of smart-application standards, describe buildings in terms of spaces, devices, sensors and observations. They are excellent at answering questions like which room a temperature reading came from, but they cannot represent the dynamic, goal-directed processes of a living metabolism: the fact that a bioreactor is digesting waste, that its microbial community is adapting, or that a design pattern borrowed from biology is governing how the system should respond to change.

The hybrid ontology bridges that gap by translating architectural metabolic design patterns, codified biological design principles drawn from the pattern-language tradition of Christopher Alexander, into a machine-readable format that AI systems can reason over. This means the home’s intelligence is not limited to monitoring data streams. Instead, the AI can interpret the state of the building’s metabolism in terms of the design intentions encoded in the ontology, adapt to dynamic biological processes that no training dataset could fully anticipate, and explain its decisions in terms that both engineers and occupants can understand. Explainability, the authors stress, is not an optional add-on but a structural requirement: because the reasoning is grounded in a shared semantic framework, the cognitive digital twin can show its work, tracing any recommendation back through the metabolic concepts and patterns that motivated it.

This approach draws on the emerging field of hybrid intelligence, which seeks to combine the complementary strengths of humans and machines rather than replacing one with the other. In the CPLS, the human is not a user issuing commands to an opaque assistant but a participant in a feedback loop that spans biology, computation and lived experience. The authors situate this within a lineage of cybernetic thinking that runs from Norbert Wiener’s control and communication theory through W. Ross Ashby’s homeostat to Gordon Pask’s conversation theory, all of which treated adaptation and dialogue between system and environment as the essence of intelligence. The home, in this view, is not a tool but a conversational partner with its own agency, what Armstrong has elsewhere described as persuadable matter.

The biological substrate is equally concrete. Microbial fuel cells, first observed by M. C. Potter in 1911 and developed intensively over the past two decades, convert the chemical energy of organic decomposition into electricity through the action of electroactive bacteria such as Shewanella oneidensis. Research teams, including collaborations involving Armstrong and ioannis Ieropoulos, have demonstrated microbial fuel cells powering lights and digital interfaces from urine and wastewater, with projects such as PEE POWER in Uganda showing real-world sanitation and energy benefits. The Living Architecture (LIAR) project built prototype metabolically engineered building units that process greywater and generate resources. The CPLS proposal scales this logic to the whole house: the gut is not a single device but a distributed digestive system woven through the architecture, managing carbon, nitrogen, water and energy flows in closed loops.

The societal implications the authors draw from this are as significant as the technology. They propose redefining comfort, the holy grail of conventional smart home design, as metabolic security: not a constant temperature and a well-stocked refrigerator, but the assurance that the household’s living systems are resilient, regenerative and productive. A metabolising home would not just reduce its environmental footprint; it would actively contribute to the household’s material economy, transforming waste into energy and nutrients, and repositioning the dwelling as a regenerative asset rather than a sink for embodied carbon and operational energy. This matters because the built environment is a major contributor to the climate crisis, both through operational emissions and through the embodied carbon of construction materials, a problem initiatives such as the European Green Deal and the World Green Building Council’s embodied carbon agenda have made impossible to ignore.

Equally transformative is the role the framework assigns to occupants. In today’s smart home, the human is largely a data source: activity recognition systems, graph neural networks and predictive energy optimisers infer what people are doing so machines can act on their behalf. The CPLS inverts this relationship. Because the ontology makes the home’s metabolic processes legible, occupants can learn to read them, to understand what their bioreactors are doing, how their energy and nutrient cycles are flowing, and where their stewardship matters. The authors describe this as transforming occupants into literate stewards, a form of domestic ecological literacy that echoes educational research on hybrid intelligence, where AI systems are designed to extend human learning rather than substitute for it. The home becomes, in effect, an instrument of ongoing education about the living world.

None of this is presented as a finished product. The paper is explicitly a conceptual contribution, a proposed foundation rather than a deployed system, and the authors are careful to note that no datasets were generated or analysed in the study itself. Its original contribution lies in the synthesis: metabolic architecture, cybernetics, semantic web technologies and human-centred AI are brought together into a single coherent framework with a methodology for hybrid human-AI systems. But the individual components are real, from validated microbial fuel cell technology to mature ontological standards and cognitive digital twin research in domains from healthcare to energy management. The wager of the paper is that the future of sustainable dwelling may lie not in smarter gadgets but in cultivating living architectural metabolisms, homes that participate in the regenerative logic of the living world rather than standing apart from it. If that wager pays off, the smart home of the coming decades may look less like a remote control with a roof and more like something closer to a body: fed by waste, warmed by microbes, guided by an explainable mind, and inhabited by people who finally understand that they are not the owners of a machine but partners in a metabolism.

Subject of Research: A hybrid ontology linking metabolic architecture and human-centred AI for regenerative smart homes

Article Title: A hybrid ontology for metabolic architecture and human-centred AI: an explainable new vision for the home

Article References: Armstrong, R., & Moodley, D. (2026). A hybrid ontology for metabolic architecture and human-centred AI: an explainable new vision for the home. AI & SOCIETY. https://doi.org/10.1007/s00146-026-03301-5

Image Credits: AI Generated

DOI: 10.1007/s00146-026-03301-5

Keywords: metabolic architecture, hybrid ontology, human-centred AI, Cyberphysical Living System, cognitive digital twin, regenerative design, smart homes, microbial fuel cells, semantic web, explainable AI, sustainability, built environment