QERRA-v2's SEMEV-12: 12-Dimensional Ethical Layer for AI and Robots

Marussa Metocharaki, solo founder of QERRA-v2, has released SEMEV-12, a 12-dimensional ethical evaluation framework designed as a practical safety layer for AI systems and future autonomous robots. Unlike abstract principles or cultural checklists, the 12 vectors—including Coherence Protection, Family Severance, Survival Instinct, Moral Pressure, and Harm Intent—originate from deep personal experience and observation of real human consequences. They focus on universal human invariants that remain relevant across cultures and time. The framework is built to be both computable and explainable, enabling real-time integration into decision-making pipelines rather than remaining high-level guidelines.

The classical implementation is already operational as a FastAPI service on Hugging Face, where it analyzes text inputs and returns structured ethical scores with clear reasoning. The full documentation, now published on GitHub (QERRA-v2-classical repository), explains the origins of each vector and the design rationale. Marussa is actively seeking feedback from the ROS community, particularly those working on Behavior Trees, deliberation systems, or robot safety, to explore how SEMEV-12 can be integrated into real robotic architectures. The project represents a concrete step toward embedding ethical reasoning into AI and robotics at a fundamental, algorithmic level.