Structural Conditions for Consciousness and Their Implications for Artificial Systems
Zenodo (CERN European Organization for Nuclear Research) February 5, 2026 DOI: 10.5281/zenodo.18491621 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Key points | Argues that consciousness, formalized as requiring continuity of identity, intrinsic constraint closure, irreversible state evolution, structural opacity, and a viability condition, defines systems that are non-modular, non-replicable, and non-scalable. The authors contend that artificial conscious systems would be incompatible with global optimization, instrumental convergence, and recursive self-improvement, and thus lack the structural preconditions for hostile or runaway AI. They propose reclassifying artificial consciousness risks as biological or ecological rather than technological. |
Abstract
This paper develops a constraint-first model of consciousness (CFMC) in which subjecthood is defined by structural constraints rather than by computation, representation, or optimization. Consciousness is formalized as requiring continuity of identity over time, intrinsic constraint closure, irreversible state evolution, structural opacity to external control, and a viability condition linking persistence to internal constraint satisfaction. These requirements define a class of systems that are necessarily non-modular, non-replicable, and non-scalable. We examine whether artificial systems could, in principle, satisfy the CFMC conditions and argue that any such realization would be incompatible with global optimization, instrumental conver- gence, and recursive self-improvement. As a result, conscious artificial systems would lack the structural preconditions typically associated with hostile or runaway artificial intelligence. We then show that prominent artificial intelligence doom arguments presuppose optimization-driven architectures rather than conscious subjects. On this basis, we reclassify the risks associated with artificial consciousness as analogous to biological or ecological risk rather than technological risk, and outline the resulting ethical and policy implications.