Biogenic Security: Why We Need a New Framework for the Life We Are Creating
We still talk about “biosecurity” as if biology were a closed system. As if the only threats worth governing are natural pathogens or sloppy laboratory practices. That framework served the twentieth century well, but it is already outdated. We now produce organisms, quasi-organisms, biological machines, and bio-digital hybrids at a pace that outstrips our regulatory imagination. Some of these systems replicate, some adapt, some learn, and some show early signs of agency that are not quite life but far beyond machinery.
This is the emerging domain I call biogenic life, and the governance we need is biogenic security.
This isn’t a philosophical exercise. It is a practical response to what scientists are already doing in labs, start-ups, agricultural systems, and AI research groups. If we continue using twentieth-century categories to regulate twenty-first-century entities, we will be constantly surprised, and occasionally blindsided.
Biogenic security is not about restricting innovation. It is about understanding what we are making and preparing accordingly.
Three New Classes of Systems We Must Account For
1. Synthetic Life
These are fully biological systems built or heavily modified by humans. Engineered microbes. Designed cell lines. Lab-grown organoids that now show sensory responses and learning patterns. They reproduce, mutate, and evolve. They can escape, recombine, and surprise us. Nothing in the classical biosecurity literature prepares us for systems that are biological but not evolutionary products.
Biogenic risk: runaway adaptation, emergent traits, ecosystem interaction, unexpected sentience thresholds.
2. Emergent Collectives
This category is often misunderstood. These are not single organisms. They are systems in which biological, digital, and mechanical components interact to produce unified behaviour. Hospital microbiomes linked to algorithmic cleaning robots. Agricultural ecosystems connected to remote machine-learning feedback loops. Urban systems where sensors, pests, microbes, and automated regulation form self-stabilising cycles.
Individually, none of these components meets any definition of “life”. Collectively, they often behave like a coordinated organism that adapts and optimises with a stability no human designed.
Biogenic risk: unanticipated collective behaviour, self-reinforcing feedback loops, multi-node “organisms” that cannot be turned off.
3. Bio-Digital Hybrids
These are the entities that truly challenge current thinking. Neurons connected to neural nets. Organoid-AI training loops. Synthetic tissues embedded with onboard learning algorithms. Biological substrates that absorb machine learning feedback. These systems exhibit learning and pattern recognition. They modify themselves. They retain memory. They attempt goals.
Biogenic risk: unclear agency, moral status uncertainty, goal escalation, opaque learning processes.
The Biogenic Triad as a Governance Threshold
To cut through the confusion, we need a clear test for when a system moves into the zone where both moral status and risk profile change. The Biogenic Triad is simple and empirically measurable:
Self-production (the system maintains or replicates itself)
Self-organisation (it coordinates its components coherently)
Self-correction (it adapts or learns via feedback)
When a system ticks all three, we are no longer dealing with a passive artefact. We are dealing with a proto-agent. This does not mean it deserves rights in the human sense, but it does mean we cannot treat it as a disposable tool. Nor can we rely solely on classical containment or kill-switch logic.
The triad should be our bright line.
A Proposed Regulatory Ladder: Biogenic Security Levels (BSLs)
BSL-A: Tools
Systems with none of the triad traits. CRISPR kits that cannot self-replicate. Static circuits. Non-adaptive robotics. Current frameworks work well here.
BSL-B: Constrained Autogenesis
Systems that show one or two triad traits but remain inside hard constraints. Engineered microbes with strict nutrient dependencies. Hybrids that cannot self-modify outside defined boundaries.
Controls: licensing, containment audits, and environmental dependency locks.
BSL-C: Proto-Agents
Systems that exhibit all three traits, but without clear self-modelling. Adaptive microbial consortia. Learning organoids. Ecological-algorithmic collectives.
Controls: telemetry logs, redundancy in shutdown mechanisms, mandatory ethical oversight.
BSL-D: Moral Status Candidates
Systems that show the triad, along with indicators of preference formation, avoidance behaviour, or goal stabilisation. These systems need a precautionary approach. Not because we are sure they deserve protection, but because we cannot rule it out.
Controls: harm-minimisation protocols, consent-analog procedures, independent governance boards.
A Practical Toolkit for Engineers and Regulators
Telemetry by default: every biogenic system logs adaptation events, resource flows, and boundary breaches.
Evolution budgets: clear thresholds for how far a system is allowed to adapt before requiring relicensing.
Redundant shutdowns: biochemical plus algorithmic plus environmental methods, so no single layer becomes a failure point.
Suffering tests: if the system shows nociception, avoidance, or preference hierarchies, escalate immediately.
Liability linked to agency: the more autonomy a system exhibits, the greater the steward’s duty of care.
No single tool will be adequate. Biogenic systems are messy. That’s the point.
Two Questions We Must Start Asking
These are not rhetorical questions. They are policy-shaping questions that deserve wide public debate.
1. If a system meets the biogenic triad, what minimal protections do we owe it?
At a minimum, protection from unnecessary harm and clarity around shutdown processes.
2. Who decides whether a system meets the triad?
Regulators? Research institutions? Independent biogenic ethics boards?
This decision cannot be left to commercial actors alone.
Why This Matters Now
This decade will be remembered as the moment biology and computation fused into a continuum. We are no longer modifying life. We are creating new classes of living and near-living systems. Failing to track this shift will lead to predictable mistakes. Overreaction in some areas. Dangerous underreaction in others.
Biogenic security is simply a recognition of reality.
We are not protecting biology from threats.
We are protecting society from the consequences of creating entirely new forms of life.
The sooner we begin thinking this way, the safer the transition will be.