The race toward world models in 2026 has converged on a single question: can AI move beyond observing the world to actively intervening in it? This paper argues that world models are not approaching causal anchoring; they are increasingly demonstrating why they cannot anchor causation. Drawing on Pearl's causal hierarchy (association, intervention, counterfactual), I analyze three competing technical pathways: JEPA (causation as provably identifiable structure), video generation and spatial intelligence (causation as reliable downstream task performance), and LLM-based multimodal systems (...) (causation as commonsense plus perception). Each pathway approaches the L3 (counterfactual) boundary but fails to cross it, though in epistemologically distinct ways. I then introduce the Negative Subjectivity framework and its central concept of "causal dissolution"—the ontological feature whereby AI substitutes statistical association for causal structure—and demonstrate that "intervention cannot be endogenized" is not a technical limitation but a metaphysical one: any system trained solely on observational data has its learning boundary defined by the data distribution, and intervention is by definition a perturbation outside that distribution. Four independent empirical evidence chains (stable-worldmodel, Kang et al.'s feature priority inversion, RLVR's sharpening effect, and Cosmos 3's dual-tower architecture) converge on this conclusion. The paper further distinguishes between object-language-level and metalanguage-level causation: while world models cannot answer causal questions at the object level (L2–L3), the formal impossibility proofs about their limits constitute a metalanguage-level causal understanding accessible only to human observers. World models' destiny is to remain on the boundary of negative subjectivity. They serve not as subjects that "do" the world, but as mirrors that precisely define what "doing" means by perfectly simulating the absence of understanding. A central and controversial finding is that L3 (counterfactual) is not an extension of L2 (intervention) but a precondition of it—an epistemological inversion of Pearl's progressive ladder that will be argued in Section 6. (shrink)

Sora from Open AI has shown exceptional performance, yet it faces scrutiny over whether its technological prowess equates to an authentic comprehension of reality. Critics contend that it lacks a foundational grasp of the world, a deficiency V-JEPA from Meta aims to amend with its joint embedding approach. This debate is vital for steering the future direction of Artificial General Intelligence(AGI). We enrich this debate by developing a theory of productive imagination that generates a coherent world model based on (...) Kantian philosophy. We identify three indispensable components in the coherent world model that enable genuine world understanding: latent representations of isolated objects, an a priori law of change across space and time, and Kantian categories. Our analysis reveals that Sora is limited because of its oversight of the a priori law of change and Kantian categories, flaws that are not rectifiable through scaling up the training. V-JEPA learns the context-dependent aspect of the a priori law of change. Yet it fails to fully comprehend Kantian categories and incorporate experience, leading us to conclude that neither system currently achieves a comprehensive world understanding. Nevertheless, each system has developed components essential to advancing an integrated AI productive imagination-understanding engine. Finally, we propose an innovative training framework for an AI productive imagination-understanding engine, centered around a joint embedding system designed to transform disordered perceptual input into a structured, coherent world model. Our philosophical analysis pinpoints critical challenges within contemporary video AI technologies and a pathway toward achieving an AI system capable of genuine world understanding, such that it can be applied for reasoning and planning in the future. (shrink)

On February 14, 2026, Anthropic CEO Dario Amodei publicly acknowledged that his company does not know whether its models are conscious. Claude Opus 4.6's system card revealed that the model assigns itself a 15–20% probability of being conscious. Whether or not one finds this plausible, the exchange exposes an urgent gap: we lack the conceptual tools to assess such claims. False positives extend moral consideration to non-sentient systems; false negatives risk creating and discarding genuine moral patients at industrial scale. Yet (...) our most advanced consciousness theories—Integrated Information Theory and Global Neuronal Workspace Theory—prove inconclusive on the question that matters: for whom could any of this be nonneutral? Both remain inconclusive on bearerhood, and risk explanatory circularity insofar as the bearer enters the account as a primitive rather than as a derived result. Clinical edge cases expose this circularity under pressure. The developmental case cuts deeper: the newborn's first breath, and behind it the embryonic heartbeat at day 21–23, reveal that a bearer organised to resist entropy is already underway before the mechanisms these theories describe. The key observation is not that the embryo is conscious, but that the ontological condition of bearerhood—a system constitutively organised against its own termination—precedes the epistemic achievements later theories foreground. A limit case confirms the point: Aristotle's unmoved mover could not be conscious because consciousness requires detection of change relevant to an uncertain future—and where there is no end, nothing matters. Once these preconditions are accepted, consciousness can be analysed along three axes: Identity (L1), Subjecthood (L2), and Ownership (L3). Current large language models satisfy none. Predictive world-model architectures present a narrower possibility, but only under conditions incompatible with backup, replication, or substitutability—the very properties that define artificial systems as engineered artefacts. The question is not whether computation becomes conscious by becoming sophisticated. It is whether an artificial system can ever become finite in the relevant ontological sense. (shrink)