A cybernetic meta-structure for endogenous AI ethics: ensuring systemic integrity through feedback loops

Purpose Current approaches to AI ethics predominantly rely on exogenous regulation, imposing constraints via external supervision such as Reinforcement Learning from Human Feedback (RLHF) or static rule sets. From a cybernetic perspective, these methods lack the requisite variety to manage novel ethical perturbations and treat ethics as a boundary condition rather than an internal systemic capacity. This article proposes an alternative paradigm of endogenous AI ethics, where ethical reasoning emerges as a homeostatic property of the system’s meta-structural design. Design/methodology/approach We introduce the Ontological Meta-Structure Engine (OMSE), a second-order cybernetic architecture designed to preserve systemic integrity. Drawing on meta-ontological analysis and systems theory, OMSE implements three operational feedback mechanisms that maintain categorical coherence (Existence Preset Index), ensure reasoning consistency under input perturbation (Cognitive Pathway Stability) and guarantee the closure of accountability loops (Responsibility Closure Index). Findings We demonstrate that external ethical constraints suffer from structural limitations, including an inability to generalize to unfamiliar contexts and a lack of transparent causal tracing. By contrast, the OMSE framework enables systems to develop ethical capabilities endogenously. Complexity analysis indicates that these integrity monitoring layers introduce a modest computational overhead of approximately 5–10% while enabling verifiable, robust ethical reasoning across novel scenarios. Originality/value This article positions philosophy not merely as commentary but as generative architecture for AI design. The OMSE framework bridges abstract meta-ontology and engineering practice, treating integrity not as a virtue but as a verifiable system state. By embedding second-order cybernetic principles into artificial intelligence architecture, we enable autonomous ethical development while maintaining structural coherence, which represents a necessary condition for safe deployment of increasingly autonomous artificial agents in complex social systems.