The convergence of autonomous software agents and living materials is reshaping how makers, designers, and laboratories approach fabrication. What once was a craft practice,growing mycelium skins, brewing kombucha-derived bacterial cellulose, or programming microbes for sensing,now sits at the intersection of synthetic biology, materials science and automated decision systems.
Over the past three years, both disciplines have matured: engineered living materials (ELMs) have moved from lab curiosities to demonstrators with programmable properties, while self-driving labs and agentic AI systems have advanced closed-loop experimentation and remote execution. Together they are forcing a rewrite of the maker workflow, from informal studio practices to regulated, data-driven pipelines.
The rise of engineered living materials
Engineered living materials embed living cells,microbes, fungi, or mammalian cells,into material architectures so the living component supplies function: sensing, repair, growth and adaptation. Reviews over the past several years have crystallized the field’s core: ELMs are distinct from passive biomaterials because their biological components continue to metabolize and respond after fabrication.
Recent polymer-chemistry and materials-science studies have emphasized the importance of designing the living,material interface: small molecular tweaks to scaffolds and gene circuits can tune mechanical properties, signalling and durability, which in turn affects how a maker or manufacturer must handle and curate the living system. This molecular-to-structural linkage is a practical reason workflows are being rethought.
Concrete examples are proliferating. Work on bacterial cellulose (often sourced from kombucha SCOBYs or Komagataeibacter strains) shows programmable mechanical properties through genetic and process control, while mycelium-based composites continue to be used in architecture, packaging and design studios as biodegradable structural materials. These technical advances are enabling makers to treat living substrates as designable, application-ready media.
How autonomous agents enter the wet lab
Autonomous agents,software systems that plan, decide and act with varying degrees of autonomy,are no longer confined to code. In laboratory contexts, agents are being paired with robotic hardware and cloud-based wet labs to run experiments without continuous human steering. Recent reporting and reviews identify a trend: large language models and decision-making agents are being integrated as orchestration layers for experiment design and scheduling.
Self-driving laboratories (SDLs) implement a closed-loop design,make,test,learn cycle, where machine learning proposes conditions, robots execute protocols, and analytics update the models. The same architectural pattern that accelerated materials discovery is now being applied to living systems,but living systems add constraints: cell viability, contamination risk and long biological timescales complicate automation.
Practically, agentic systems help by lowering the cognitive load of protocol design, managing large parameter spaces, and enabling remote experimentation. Case studies in 2025,26 document LLM- and multi-agent-driven optimization of workflows such as cell-free protein synthesis and other biosynthetic protocols, illustrating how agents can propose and iterate biologically relevant experiments at speed.
Closed-loop workflows: design-build-test-learn with living systems
The classic DMTA (design,make,test,analyze) loop expands when the material itself is alive: the ‘make’ step must account for growth cycles, environmental control and potential evolution, while ‘test’ often requires biological assays rather than single-shot physical measurements. Self-driving-lab literature highlights this increased complexity and the need for different automation strategies.
For engineered living materials, feedback is not only about performance metrics (strength, conductivity) but also viability, stability and biosafety markers. Recent work on probiotic-based living therapeutics and engineered microbes stresses the importance of measuring genetic stability, horizontal gene transfer risk and immune interactions,metrics that must be integrated into autonomous decision rules when machines control experiments that involve living cells.
Technically, this means workflows now combine standard materials analytics (mechanical, optical) with microbiological readouts (CFU counts, sequencing, reporter outputs) and closed-loop optimization strategies (Bayesian optimization, active learning). The result is a hybrid pipeline where digital agents coordinate heterogeneous data streams and time-dependent experiments, forcing workflow designers to bridge lab automation, data infrastructure and biological practice.
Practical maker practices: community labs and hybrid fabrication
On the maker end, living materials remain attractive because they are low-energy, locally producible and materially rich. Design studios and community bio-labs have incorporated workshops on mycelium forming, kombucha-derived cellulose fabrication and algae-based bioplastics,practical entry points that teach embodied, hands-on techniques while exposing makers to biological temporality and care practices. HCI and design research documents this shift towards more-than-human fabrication in maker contexts.
However, when autonomous agents and remote automation enter these contexts, tensions arise. Makers accustomed to tactile feedback and slow growth cycles face new choices: to keep the craft studio model or to outsource parts of the workflow to cloud labs and agent-orchestrated instrumentation. Community labs and studio-practice literature show hybrid models emerging,studio prototypes that are grown locally but characterized and optimized via remote, agent-driven analysis.
These hybrid paths lower barriers for non-experts to experiment with living materials while raising questions about reproducibility, intellectual property and custodianship of living strains. Makers adopting agent-supported workflows must learn to interpret digital recommendations, maintain rigorous records, and manage material provenance in ways that mirror small-scale biomanufacturing.
Regulatory, ethical and biosafety considerations
Embedding living systems in consumer-facing objects and using autonomous agents to manipulate them triggers regulatory scrutiny. Regulators in major jurisdictions are actively re-evaluating frameworks to account for novel biological products and hybrid materials; policy analyses note that classification and oversight for ELMs and living therapeutics is still evolving. Makers and startups must therefore navigate FDA, EPA, EMA and EU rulemaking that touches on genetically modified organisms, medical products and novel food/material categories.
Biosecurity and biosafety are central concerns when agents design experiments remotely. Reviews of self-driving labs and autonomous experimentation emphasize safety layers: robust access controls, provenance tracking, anomaly detection and human-in-the-loop gates for high-risk steps. For living materials, additional safeguards include containment, strain registries, and assays for genetic stability,requirements that reshape simple maker recipes into regulated processes when scaled or distributed.
Ethically, practices must account for stewardship of living systems and community consent. HCI studies in bioart and community bio spaces highlight how public engagement, transparent protocols and education can reduce harm and build accountable practices when living materials become visible in everyday objects and exhibitions. Policymakers and platform builders are therefore being asked to design governance that balances innovation and precaution.
From studio to factory: scalability and industry adoption
Translating maker-scale living materials into industrial products requires new workflows: controlled bioreactors, downstream processing, quality control and supply-chain integration. Recent materials-chemistry perspectives argue that the living,polymer interface and process standardization are the technical bottlenecks for scaling. Advances in genome editing of production strains and eco-friendly process innovations are helping, but many questions remain about consistency and cost.
Market signals are mixed. Some commercial ventures in mycelium leather and other ELMs have shown traction, but there have also been business failures and restructurings,reminders that scaling biological production entails capital-intensive infrastructure and regulatory pathways. These market dynamics are important context for makers considering commercial ambitions or partnerships with larger manufacturers.
At the same time, modular self-driving-lab platforms and lower-cost automation projects seek to democratize scale-up by providing cloud or shared infrastructure where makers and startups can validate processes before committing to full-scale production. Such shared pathways may become the de facto route from studio prototypes to regulated products if they combine robust biosafety, transparent data and industrial QA standards.
The encounter between agentic AI and living materials is rewriting workflows at multiple levels: technical (digital+biological integration), institutional (new governance and infrastructure), and cultural (new maker literacies and ethics). For professionals and policymakers, the imperative is to build interoperable systems that embed safety, provenance and robust analytics without stifling creative experimentation.
For practitioners, the practical takeaway is simple: adopt data-centric practices, treat biological affordances as first-class design constraints, and plan for regulatory and safety gates early. The future of making with living materials will be neither purely artisanal nor purely automated; it will be a hybrid ecosystem where agents and organisms co-author materials and where workflows are intentionally designed to manage risk, value and creativity.
