paper
active
2022
ⓒ 28
paper:generalizing-frameworks-for-sentience-beyond-natural-species

Generalizing frameworks for sentience beyond natural species

ByMichael LevinⓘAllen Discovery Center At Tufts University, Allen Discovery Center, Tufts University + 10 more
DOI 10.51291/2377-7478.1733OpenAlex W4285221823
Basal Cognition FrameworkExpanded cognition & sentience frameworksSubstrate-independent sentience frameworksSentience Detection FrameworkCyborg AgentsCrump et al. Eight Criteria FrameworkLaMDA (Language Model for Dialogue Applications)Reducing false negatives in sentience attribution"We stand at the entry point to a great transition, in which all of Darwin's familiar 'Forms Most Beautiful' are just a tiny region of an astronomically large space of possible beings."

TL;DR

Criteria anchored to vertebrate neuroanatomy and verbal behavior — exemplified by the Smith & Boyd (1991) framework and the Turing Test — are structurally inadequate for the full space of possible sentient agents, and the Crump et al. (2022) 8-criterion decapod framework, while a genuine advance, must now be generalized far beyond natural phylogenetic lineages. Levin argues that associative learning, Crump et al.'s criterion #7, already occurs in gene regulatory networks and non-neural morphogenetic agents, meaning the pivot from 'neurons' to 'electrically active cell' unlocks most of the 8 criteria for substrates including organoids, xenobots (Kriegman et al., 2020, PNAS 117:1853), cultured-neuron 'hybrots' that learned in simulated game-worlds (Kagan et al., 2021), and synthetic living machines built on frog-cell platforms (Blackiston et al., 2021, Sci Robot 6:eabf1571). The instrument Levin implicitly introduces is a substrate-neutral invariant search: replacing phylogenetic or anatomical heuristics with deep functional invariants — such as 'competency in navigating arbitrary problem spaces' (Fields & Levin, 2022) — that apply across biological, chimeric, AI, and exobiological agents alike. The LaMDA debate (Thopilian et al., 2022) illustrates the cost of this gap: strong opinions circulate with no defensible criteria. Levin argues this implies that developing principled, quantitative, substrate-independent sentience frameworks is an existential requirement for humankind, not merely an academic refinement, because the coming decades will embed genuinely novel agents — cyborgs, biorobots, AI systems, synthetic organisms — into society before ethics has caught up.

What to take away

  1. 1. The Smith & Boyd (1991) sentience criteria, designed for biomedical animal research, are irrelevant across the vast majority of the space of possible agents because they rely on vertebrate neuroanatomy and phylogenetic provenance as proxies.
  2. 2. Crump et al.'s (2022) criterion #7, associative learning, is already demonstrable in gene regulatory networks (Biswas et al., 2021, iScience 24:102131; Fernando et al., 2009, J R Soc Interface 6:463), meaning non-neural substrates satisfy at least this sentience indicator.
  3. 3. Cultured-neuron 'hybrot' systems — networks of biological neurons embodied in robotic bodies — exhibited learning and proto-cognitive behavior in closed-loop platforms including a simulated game-world (Kagan et al., 2021, bioRxiv 2021.12.02.471005), making substrate-neutral sentience assessment practically urgent.
  4. 4. The xenobot lineage (Kriegman et al., 2020, PNAS 117:1853; Blackiston et al., 2021, Sci Robot 6:eabf1571) demonstrates coherent, autonomous synthetic living machines with anatomies radically unlike any evolved organism, directly challenging sentience frameworks that require homology to known life.
  5. 5. A methodology another researcher could replicate is the 'simple pivot' Levin proposes: systematically substituting 'electrically active cell' for 'neuron' across each of Crump et al.'s 8 criteria and testing whether non-neural morphogenetic agents (e.g., planaria, organoids) satisfy them empirically.
  6. 6. The LaMDA large language model debate (Thopilian et al., 2022, arXiv:2201.08239) is cited as a live case where neither Turing-Test-like verbal performance nor brain-homology criteria suffice to resolve sentience attribution, leaving the field without a defensible decision procedure.
  7. 7. Human perceptual agency-detection is well-tuned for medium-sized objects moving at medium speeds in 3D space but is not adapted to recognizing intelligence in unfamiliar problem spaces such as physiological, metabolic, transcriptional, or anatomical domains (Fields & Levin, 2022, Entropy 24).
  8. 8. An open hypothesis raised is whether Crump et al.'s criterion #8 (analgesia preference, with three sub-criteria: self-administration, location preference, and prioritization) can be operationalized and tested in non-neural morphogenetic agents — work described as currently underway but not yet reported.
  9. 9. Ontogenetic continuity — the gradual transition from a quiescent oocyte ('just physics') to a human-level mind — is used as a second axis alongside phylogeny to argue that sentience thresholds cannot be binary and must be treated as continuous, with N=1 Earth phylogeny being an insufficient reference set.
  10. 10. Levin predicts that failing to develop substrate-neutral, quantitative sentience frameworks before cyborgs, biorobots, AI agents, and synthetic organisms are embedded in homes and bodies constitutes an existential ethical risk, not merely a theoretical gap.

Peer brief — for seminar discussion

Michael Levin's 2022 commentary in Animal Sentience (32:15, DOI 10.51291/2377-7478.1733) uses Crump et al.'s 8-criterion decapod sentience framework as a launching point to argue that all existing sentience-assessment tools — including the Smith & Boyd (1991) biomedical criteria and Turing-Test-style verbal behavior — are structurally inadequate for the rapidly expanding space of non-natural agents. The paper surveys four converging technological fronts — synthetic morphology (xenobots: Kriegman et al., 2020, PNAS 117:1853; Blackiston et al., 2021, Sci Robot 6:eabf1571), hybrot systems (Kagan et al., 2021, bioRxiv 2021.12.02.471005), large language models (LaMDA: Thopilian et al., 2022, arXiv:2201.08239), and chimeric biorobotics — to establish that beings with no phylogenetic touchstone and no neuroanatomical homology to vertebrates already exist in laboratories and will soon enter society. The load-bearing finding is that most of Crump et al.'s 8 criteria, specifically including criterion #7 (associative learning), are already satisfied by non-neural substrates such as gene regulatory networks (Biswas et al., 2021, iScience 24:102131) and morphogenetic tissues, meaning the criteria are substantively broader than their biological framing implies. The method Levin introduces is a substrate-neutral invariant search: replacing phylogenetic and anatomical proxies with deep functional invariants — particularly 'competency in navigating arbitrary problem spaces' as formalized in Fields & Levin (2022, Entropy 24) — that apply equally to biological, chimeric, AI, and hypothetical exobiological agents. An alternative method the commentary implicitly sets aside is the integrated information theory (IIT) approach, which offers substrate-independence but is not engaged here, likely because its first-person grounding faces the same Hard Problem Levin explicitly brackets. The implication Levin draws is normative and urgent: developing principled, quantitative, substrate-independent sentience frameworks is an existential requirement for humankind, because ethical frameworks will otherwise lag behind deployment of novel agents. The paper's explicit prediction is that the coming decades will introduce agents into homes and bodies that offer none of the familiar phylogenetic or anatomical heuristics humans have historically used. A critical reader would push back most forcefully on the evidential weight Levin places on the xenobot and hybrot systems as sentience-relevant cases. Demonstrating associative learning or closed-loop adaptive behavior in cultured neurons or frog-cell assemblies does not establish that these systems have morally relevant welfare states — it may only show that the behavioral criteria in Crump et al. were never sufficient proxies for sentience in the first place, which is a different claim than that these substrates are sentient. Levin acknowledges he is not solving the Hard Problem, but the move from 'these systems satisfy functional criteria' to 'we therefore have potential moral obligations toward them' inherits precisely the gap he sets aside. The commentary is also, by design, a programmatic call rather than an empirical contribution, so it offers no new data and no worked example of the substrate-neutral framework applied to a specific novel agent — a gap that limits its immediate operationalizability for researchers wanting to act on its recommendations.

Frameworks (1)

Findings (2)

Claims (25)

Hypotheses (3)

Questions (6)

Original abstract (expand)

Crump et al. (2022) offer a well-argued example of an essential development: a rigorous framework for assessing sentience from the perspective of moral concern over an agent’s welfare. Current and forthcoming developments in bioengineering, synthetic morphology, artificial intelligence, biorobotics, and exobiology necessitate an expansion and generalization of this effort. Verbal reports (the Turing Test) and homology to human brains are utterly inadequate criteria for assessing the status of novel, unconventional agents that offer no familiar touchstone of phylogeny or anatomy. We must develop principled approaches to evaluating the sentience of (and thus, our responsibility to) beings of unfamiliar provenance and composition.

Related work— refs + corpus + external arXiv

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

+22 more

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
    What are the appropriate criteria for assessing sentience in novel, unconventional agents that lack familiar phylogenetic origins or anatomical structures?questions/what-are-the-appropriate-criteria-for-assessing-sentience.md0.850
  • aboutblank_kb
    Levin, Michael (2022) Generalizing frameworks for sentience beyond naturalpapers/linkified/levin-michael-2022-generalizing-frameworks-for-sentience-beyond-natural.md0.849
  • aboutblank_kb
    Levin, Michael (2022) Generalizing frameworks for sentience beyond naturalpapers/edited/Generalizing frameworks for sentience beyond natural species_edited.md0.849
  • aboutblank_kb
    Levin, Michael (2022) Generalizing frameworks for sentience beyond naturalpapers/cleaned/Generalizing frameworks for sentience beyond natural species.md0.844
  • aboutblank_kb
    Can synthetic living machines exhibit genuine cognition and intentional behavior without evolved neural systems?questions/can-synthetic-living-machines-exhibit-genuine-cognition-and.md0.841
  • aboutblank_kb
    Crump Et Al.'S Eight Criteria For Sentienceframeworks/crump-et-als-eight-criteria-for-sentience.md0.832
  • aboutblank_kb
    Can artificial systems designed to be living achieve comparable cognition and adaptability to natural organisms?questions/can-artificial-systems-designed-to-be-living-achieve.md0.831
  • aboutblank_kb
    Can xenobots and other synthetic organisms reveal principles of biological agency and minimal intelligence?questions/can-xenobots-and-other-synthetic-organisms-reveal-principles.md0.827
  • aboutblank_kb
    Smith & Boyd Criteriaframeworks/smith--boyd-criteria.md0.826
  • aboutblank_kb
    Is developing principled frameworks for recognizing sentience in diverse intelligences an existential requirement for humanity?questions/is-developing-principled-frameworks-for-recognizing-sentience-in.md0.820
  • aboutblank_kb
    How can synthetic organisms with alternative morphologies or substrates achieve goal-directed adaptive behavior?questions/how-can-synthetic-organisms-with-alternative-morphologies-or.md0.820
  • aboutblank_kb
    What ethical frameworks should govern the creation and use of synthetic living organisms?questions/what-ethical-frameworks-should-govern-the-creation-and.md0.819