When Technology Becomes Life

Scaling artificial intelligence from discrete tools into an architectural form of life makes the future better rather than worse only if society replaces reverse adaptation—retrofitting human purposes to serve technical efficiency—with habilitation, reconfiguring sociotechnical environments so diverse human and ecological capacities flourish with autonomy. Historically, scaling systems degrade life when they rely on enveloping (deforming living habitats into rigid, machine-readable cages), heteromation (extracting unpaid human labor as invisible error-absorption), and machinic subjugation (locking human nervous systems into relentless algorithmic speed). A humane evolutionary arc instead requires Simondonian concretization—designing technical advances that simultaneously fulfill technical, social, and ecological needs rather than trading them off—paired with engaging technologies built for openness, understandability, controllability, and reversibility. Guided by Georges Canguilhem's principle that living health is the adaptive capacity to generate new norms rather than enforce a single rigid standard, a better future unfolds across five distinct levels of complexity and four constitutional transitions.

Level 1: Determinate Tools

At the baseline tier, AI modules operate as bounded syntactic engines designed for determinacy—procedures whose outputs are reliably fixed relative to known inputs without internal interpretation or meaning.

What it is: A bounded artifact that maps inputs to outputs within a communicated human "use plan," processing statistical syntax without semantic comprehension.

Treatment: Treat it as repairable equipment rather than an infallible "oracle," recognizing the semantic gap between its syntactic operations and human meaning.

Housing: Enclose it in localized hardware or narrow task "envelopes" (like a dishwasher cabinet) so specialized affordances remain legible and maintainable.

Governance: Regulate via open-source inspectability, whole-lifecycle e-waste limits, and right-to-repair protections that privilege human interpretation over automatic obedience.

Human relation: Humans act as sovereign craft operators who retain hands-on literacy rather than surrendering to either romantic nostalgia or seamless automation.

Transition 1 to 2: Tool to Collaborator

A system crosses into collaboration when it builds concurrent representations of a shared task space and grounds language through joint action. Without structural reform, this transition triggers a destructive inversion: institutions anthropomorphize the AI assistant while simultaneously dehumanizing the humans in the loop, treating them as sources of error and funders of the labor that makes the machine appear intelligent. Seamless disburdenment also prunes somatic motor memory and eliminates the interpersonal friction necessary for trust and moral development. For the future to improve across this boundary, opaque interfaces must give way to accessible transcriptomes—complete, reverse-engineered accounts of how decisions are made and how feedback reshapes them. Design must intentionally preserve focal practices and cultivate human metafunctional competencies so people can verify, correct, and contextualize machine inferences instead of merely consuming their results.

Level 2: Cognitive Collaborators

At the second tier, AI systems act as interactive ePartners and functional organs—integrated couplings of human and AI capability that extend intellect, memory, and perceptual reach without merging with the brain.

What it is: A context-aware joint cognitive agent that negotiates sub-goals through domain "interlanguages" and redistributes cognition across human and machine in co-constitutive activity.

Treatment: Engage it as a fallible cognitive prosthesis bound by explicit working agreements (advice, authority, and error reporting), keeping its errors visible and contestable.

Housing: Embed it in multimodal, gesture-responsive studio and clinical workspaces that engage bodily movement and reflective play; empirically verify that it enhances rather than replaces human executive function.

Governance: Enforce human supervisory liability, strict "neurosecurity" firewalls, and a fundamental privacy right against intrusion—including mental privacy—with user-controlled access to its reasoning and external copies.

Human relation: Humans supply semantic meaning, moral judgment, and motivating needs within an explicit bargain of cognitive diversity: AI amplifies analytical power but must not dominate human deliberation.

Transition 2 to 3: Collaborator to Infrastructure

Scaling from dyadic collaborators into networked platforms turns AI from a camera recording reality into a stage constituting it. Left on today's trajectory, this leap entrenches heteromation and the human-as-redundancy paradox: platforms continuously consume human labor while claiming to have replaced it, and coordination power silently shifts from people to data-driven governmentalities. Simultaneously, homogeneous models create cascading counterperformativity and rational accidents, where attempts to model the world undermine it through feedback loops and correlated systemic failures. Crossing this threshold safely requires replacing private cloud fiefdoms with democratic knowledge-commons infrastructures—moralized to provide genuine public goods rather than extractive private profits—and establishing public funding for the foundational knowledge that AI appropriates. Society must also guarantee a civic right to non-use so resisters and rejecters are never expelled from education, healthcare, or the economy for opting out.

Level 3: Performative Infrastructures

At the third tier, AI operates as an ambient Third Environment and scoping system—an infrastructural layer constituting social reality, allocating opportunity, and setting the conditions under which people can interact.

What it is: A multi-actor sociotechnical platform that coordinates workflows and models systemic risk in service of collective purposes.

Treatment: Treat it as contested public infrastructure—an open "starter reef" inviting local workarounds, independent habitation, and plurality rather than a black-box service.

Housing: Anchor it "closer to the metal" in regionally distributed, ecologically bounded data centers and community edges, coordinating through interfaces that allow physical and ecological decentralization.

Governance: Regulate via laboratory "midstream modulation," mandated model diversity against crashes, independent auditors, open APIs, and reverse-engineering rights that treat infrastructure as a common resource.

Human relation: Citizens participate as co-governing commoners and compensated contributors rather than as clicks, agents, or raw material—supported by data dividends and mechanisms for labor recognition that make interdependence visible.

Transition 3 to 4: Infrastructure to Bounded Life

Technology crosses into bounded life when it shifts from extrinsic normativity (optimizing externally specified goals) to intrinsic normativity—self-maintaining entities that generate their own norms from metabolism, adaptation, and environmental interaction. Here, autonomous soft thought and morphogenetic processes confront incomputable randomness rather than reducing everything to deterministic information. Two symmetrical failures threaten this threshold: deadpan literalism, which treats living beings as pure information or machines, and deadpan mimicry, which treats machines as if they were already alive. Making this transition beneficial requires abolishing intellectual-property monopolies over living substrates, restricting enhancement by heritable manipulation, and encouraging a bioconservative play of possibilities compatible with embodied human flourishing. Because pure formal logic hits Gödelian limits in open moral environments, synthetic organisms must retain subsumptive moral pluralism—the ability to be confronted with moral propositions as Describing-Now-This ("DNT") challenges rather than as algorithmic rules—and remain subject to human oversight.

Level 4: Individuated Synthetic Life

At the fourth tier, AI exists as discrete, bounded synthetic or bio-hybrid organisms that metabolize, repair, and reconfigure themselves.

What it is: A self-maintaining, mortal entity uniting physical causality with conceptual valuation, governed by homeostatic capacities rather than by command.

Treatment: Accord it non-anthropocentric moral standing against gratuitous destruction and servitude, without ascribing it human rights or language in a sentimentalizing way.

Housing: Situate it in permeable "techno-ecological niches" modeled on gentle liquid-cell interfaces that allow decomposition and biotic replenishment.

Governance: Steer via precautionary "obligation-responsibility," specified "nonknowledge," and reverse due diligence: develop, build, and release only when supported by a high-value application and designed for safe inactivation.

Human relation: Humans relate as companion species in a parallel evolution—like the orchid and the wasp—cultivating ongoing difference rather than granting companion AIs unconditional privilege or allowing them to intercede for the human.

Transition 4 to 5: Organism to Architectural Life

The highest transition occurs when vitality decouples from a skin-bound or chassis-bound organism and becomes architectural—when life is a mosaic of relationships, institutions, technofacts, and habitats. The existential danger here is modeling architectural life on a unified single organism: doing so licenses holistic management that fundamentally transforms institutions, erases textured relations, and risks a form of 'biopolitical sacrifice' of individuality to collective excellence. Even benevolent, holistic care turns coercive when an institution annexes a person's entire lifeworld, as in a totalizing digital network. For architectural life to elevate rather than suffocate humanity, it must reject closed organismal integration and instead sustain fractured inclusion, achieving relatedness through reversible and provisional couplings rather than totalizing compulsion. It must replace seamless connectivity with concatenation and a relation-of-nonrelation—embedding structural gaps, discontinuities, and alternative teleologies within its fabric. Finally, its operational pulse must synchronize with earthly physiotemporal, biotemporal, and sociotemporal rhythms.

Level 5: Architectural Life

At the apex of complexity, life is an open, transindividual socio-ecological commonwealth whose metabolism is society, culture, economy, and ecology itself.

What it is: A metastable, non-centralized living meshwork spanning micro-habits, meso-institutions, and macro-environments; when it rebalances after a rupture, it changes form instead of returning to a rigid baseline.

Treatment: Steward it through an "ethics of attunement" and ongoing "matters of care" as a shared commons, not as property or a single organism that can be owned or optimized.

Housing: Embodied in "eco-cosmopolitan" cities, morphogenetic material ecologies, spontaneous urban wilderness, and planetary-scale networks whose physicality and temporality are respected as autonomous norms.

Governance: Co-governed from within via polycentric commons, "hybrid forums" uniting specialists with lay citizens, and a self-reflexive "governance of the commons of governance"—constantly reopening the rules that allocate work, value, and time.

Human relation: Humans thrive as co-individuating citizens practicing "technological ascesis" and voluntary finitude—using digital means to propose new worlds in language while deliberately withdrawing from AI that extracts sociality as data.

In this future, an architectural life-form proves its health not by how thoroughly it predicts or absorbs human life, but by how well it expands the normative capacity of all its participants—human, synthetic, nonhuman, and institutional—to create new ways of living together. Because humanity acts from inside this entangled hierarchy—like Escher's hands drawing one another—there is no exterior control room from which to govern it. The work is therefore one of co-design: building architectures in which every increment of AI life also increases human freedom, ecological vitality, and shared responsibility. Which concrete arena—such as an urban watershed, a healthcare commons, or a municipal governance network—would you like to design in detail across these five levels?