Abstract HEDR (Hybrid Ethical-Decision Reasoning) Framework, version 2026.2, represents a substantial shift from a philosophical protocol to an implementable tool for ethical decision-making for humans and AI systems. Compared to the original Zenodo publication from 2025, this version introduces seven key innovation areas identified through extensive empirical testing on twelve real-world dilemmas of varying severity and context. Key contributions of version 2026.2: enhanced protection of vulnerable decision-makers including psychological state detection, formalization of a fourth type of reality constraint for verbal and undocumentable information sources, explicit delimitation of competencies of individual protocol steps preventing hallucination of decision paths, new jurisdictional architecture eliminating western- centric bias, and formalization of the alibism prohibition as an explicitly forbidden protocol output. Empirical testing confirmed the protocol's functionality across managerial, political, military, personal and existential dilemmas, including crisis situations with extreme time pressure. HEDR 2026.2 is consistent with the requirements of EU AI Act, GDPR, NIST AI RMF, OECD AI Principles, UNESCO 2021 and ISO 42001.
Michal Révay (Sun,) studied this question.