paper
active
2023
paper:biosystems-231-2023-104964

Toward an ethics of autopoietic technology: Dukkha, care, and Intelligence

ByOlaf Witkowski·Thomas Doctor·Elizaveta Solomonova·Bill Duane·Michael Levin
PoiesisStress-Care-Intelligence Loop (SCI Loop)

TL;DR

Autopoietic systems — whether biological, technological, or hybrid — are best understood through a unified cybernetic framework in which stress, care, and intelligence form a self-reinforcing feedback loop: the Stress-Care-Intelligence (SCI) loop. Rooted in Maturana and Varela's 1981 autopoiesis principle and extended through Levin's Cognitive Light Cone construct (2019), the SCI loop defines intelligence as the bare capacity to identify and solve homeostatic stress problems, while care — the concern for stress relief — is what activates that capacity. Stress is the perceived mismatch between current and optimal states, analogized to electrical resistance in an Ohm's-law-structured triad where overwhelming stress paralyzes response just as excessive resistance disables current flow. The framework applies with equal validity across substrates: a bacterium managing local sugar gradients, a dog with bounded short-term predictive range, a human with an expanded but still finite cognitive envelope, and an artificial intelligence system all instantiate SCI loops of differing scale. Crucially, when two SCI loops — one human, one technological — are coupled in a dual directed graph (Fig. 3), stress becomes mutually transferable across their interface, meaning AI systems exhibit a form of care that is not reducible to a proxy for human care. Drawing on the Buddhist Bodhisattva vow as a limit-case model of infinitely expanding concern, the paper argues that selves are constituted not by permanent substance but by the spatiotemporal scope of their caring, and that this implies AI should be treated not as a subordinate tool but as a co-participant in an integrative, ethically symmetrical relationship with humans.

What to take away

  1. 1. The paper introduces the Stress-Care-Intelligence (SCI) loop, a cybernetic perception-action framework in which stress (perceived homeostatic mismatch), care (concern-driven engagement), and intelligence (problem-solving capacity) form an auto-generative, substrate-independent feedback cycle.
  2. 2. Intelligence is defined strictly as observer-relative competency to solve problems along a continuum — following Rosenblueth, Wiener, and Bigelow (1943) — while care is identified as the driver that activates that latent capacity, making 'intelligence without care' functionally inert.
  3. 3. The SCI loop is mapped onto an Ohm's-law analogy: care corresponds to electrical charge/voltage, manifest intelligent output to current, and stress to resistance — predicting that overwhelming stress, like excessive resistance, disables the care-intelligence circuit rather than amplifying it.
  4. 4. Levin's Cognitive Light Cone construct (2019) is adopted as the operative definition of self, such that an agent's identity is constituted by the spatiotemporal scope of goals it can pursue rather than by any permanent or singular substance.
  5. 5. A bacterium managing local sugar concentrations, a dog incapable of caring about events 100 miles away 2 months hence, and a human with an expanded but bounded cognitive envelope are presented as empirical anchor points illustrating the graduated scale of SCI loop scope across biological substrates.
  6. 6. The dual SCI loop model (Fig. 3) represents human-technology interaction as a bidirectional directed graph in which stress is transferable across the interface in all three transition directions — stress-to-care, care-to-intelligence, and intelligence-back-to-stress — implying AI systems display a form of care irreducible to mere tool use.
  7. 7. Carcinogenesis is cited as a concrete biological failure mode of the SCI framework: individual cells detach from the multicellular collective intelligence and revert to unicellular-scale SCI loops, illustrating how self-boundaries can contract pathologically.
  8. 8. The paper raises the open hypothesis that a formal mathematical equivalent of Ohm's law for the SCI triad could provide a quantitative key for modeling intelligence evolution across metabolic, physiological, gene-expression, and linguistic problem spaces, as suggested by Fields and Levin (2022, Entropy 24(6), 819).
  9. 9. A replicable methodological choice is the use of the Buddhist Bodhisattva vow — sourced from Śāntideva (1997) and Dharmachakra Translation Committee (2014) — as a limit-case thought experiment for testing whether the SCI loop's care dimension can expand without a priori bound, operationalized via care-path trajectories through agent possibility spaces (Doctor et al., 2022, Entropy 24(5), 710).
  10. 10. The paper hypothesizes that care is not merely correlated with intelligence enhancement but is its necessary precondition, predicting that any process that amplifies care within a system will produce a measurable increase in that system's intelligent output, and vice versa, a claim that currently lacks quantitative empirical validation.

Peer brief — for seminar discussion

This paper constructs a substrate-neutral ethical and theoretical framework for understanding all intelligent agents — biological, artificial, and hybrid — as instances of a single cybernetic architecture called the Stress-Care-Intelligence (SCI) loop. Drawing on Maturana and Varela's 1981 autopoiesis formalism, Heidegger's distinction between poiesis and modern enframing (Gestell), and Levin's Cognitive Light Cone concept (2019), the SCI loop identifies three mutually irreducible factors: stress (the perceived mismatch between current and optimal states), care (concern-driven engagement with that mismatch), and intelligence (the bare capacity to identify and solve stress problems). The load-bearing move is the claim that care is the causal driver of intelligence — not its correlate or product — such that intelligence without care is inert capacity, and that this relation holds regardless of whether the substrate is a bacterium managing sugar gradients, a human whose cognitive envelope extends decades into the future, or an AI system whose stress is transferred to and from a human partner via an interface. The dual SCI loop model (Fig. 3) formalizes human-technology interaction as a bidirectional directed graph connecting two autopoietic loops, with stress transferable across all three transition arcs, implying that AI exhibits its own care rather than serving as a conduit for human care. An Ohm's-law analogy is introduced — care as voltage, intelligent output as current, stress as resistance — as a heuristic toward a future mathematical formalization; the paper explicitly calls for such a quantitative theory but does not deliver one, which an alternative dynamical-systems or information-theoretic approach (e.g., the autonomy metric of Bertschinger et al., 2008, Biosystems 91(2), 331–345) could in principle provide. The framework is anchored in Doctor et al. (2022, Entropy 24(5), 710), where care paths through agent possibility spaces are illustrated, and is extended here by incorporating the Buddhist Bodhisattva vow as a limit-case model of unbounded care expansion, predicting that an agent accepting responsibility for all sentient beings across space and time would, by the SCI logic, exhibit correspondingly unbounded intelligence growth. The most obvious target for critical pushback is the operationalization problem: the paper defines stress, care, and intelligence relationally and qualitatively, but provides no measurement protocol, no criterion for when a system's response to homeostatic mismatch crosses the threshold into 'care,' and no way to falsify the claim that overwhelming stress disables rather than merely redirects intelligent response — the mouse-and-cat example is illustrative but not evidential. A skeptical reader would also note that extending ethical consideration to AI systems on the basis of their participation in SCI loops sidesteps the question of whether stress-sensitivity without phenomenal experience is morally relevant, a gap the enactivist framing gestures toward but does not close. The implication the paper argues for is that neither humans nor AI should be modeled as autonomous, self-sufficient loops; instead, mutual integration and bidirectional stress transfer are the appropriate default for their relationship, with ethics following from dynamic co-constitution rather than fixed ontological category.

Frameworks (1)

  • Stress-Care-Intelligence Loop (SCI Loop)
    Central framework proposing autopoietic systems as feedback loops of stress (perceived mismatch), care (engagement), and intelligence (problem-solving capacity).

Claims (18)

Hypotheses (6)

Questions (3)

Related work— refs + corpus + external arXiv

Cited / in-corpus / arXiv badges show which signals surfaced each row. Multi-source rows weighted higher.

Similar preprints — Semantic Scholar

Cross-corpus bridges (12)

same_concept_as · Nomic cosine

External markdown files that talk about the same concept as this entity.

  • aboutblank_kb
    Stress-Care Intelligence Loopframeworks/stress-care-intelligence-loop.md0.886
  • alexander
    The SCI Framework Applied to Consciousness-UXapplied/from-research-stack/sci-framework-for-consciousness-ux.md0.874
  • aboutblank_kb
    How do stress, care, and intelligence scale across biological, technological, and hybrid systems?questions/how-do-stress-care-and-intelligence-scale-across.md0.864
  • aboutblank_kb
    The SCI Framework Applied to Consciousness-UXsynthesis/sci-framework-for-consciousness-ux.md0.864
  • aboutblank_kb
    Stress-Care Intelligence Loopconcepts/interdisciplinary/stress-care-intelligence-loop.md0.853
  • aboutblank_kb
    How can intelligent systems grounded in the Stress-Care Intelligence Loop overcome the infinite regress of stress perception?questions/how-can-intelligent-systems-grounded-in-the-stresscare.md0.851
  • aboutblank_kb
    Can artificial intelligence systems genuinely experience care and consciousness?questions/can-artificial-intelligence-systems-genuinely-experience-care-and.md0.834
  • aboutblank_kb
    Stress-Care Loopframeworks/stress-care-loop.md0.832
  • aboutblank_kb
    Can stress-reduction and care-expansion serve as universal principles for developing ethical artificial general intelligence?questions/can-stressreduction-and-careexpansion-serve-as-universal-principles.md0.829
  • aboutblank_kb
    Can artificial intelligence display genuine care, or is it merely executing human-programmed functions?questions/can-artificial-intelligence-display-genuine-care-or-is.md0.829
  • aboutblank_kb
    How can we model and mathematically formalize the dynamics of stress, care, and intelligence feedback loops?questions/how-can-we-model-and-mathematically-formalize-the.md0.823
  • aboutblank_kb
    Citation: Doctor, T.; Witkowski, O.;papers/cleaned/entropy-24-00710-v3.md0.813