THE RELATIONSHIP BETWEEN TRUSTED ARTIFICIAL INTELLIGENCE AND XAI 2.0: THEORY AND FRAMEWORKS

The paper investigates the relationship between Trusted Artificial Intelligence and Explainable AI 2.0 (XAI 2.0). It demonstrates that XAI 2.0 is the structural core of Trusted AI and a prerequisite for Responsible AI. A critical review of the evolution from XAI 1.0 to XAI 2.0 confirms the need for multilayer, contextadaptive explanations. A neurosymbolic hierarchical architecture is proposed, integrating deep networks with fuzzy layers, ontological reasoning, and an LLMbased verbalisation module. A comparative study of leading frameworks and guidelines reveals convergence of explainability practices throughout the MLSecOps pipeline. A sevenlevel explainability scale (L0–L6) and an integral Trust Index are introduced to formalise progress toward selfexplanatory intelligence (L6). The solution already achieves levels L4–L5 and provides an engineering foundation for selfauditing, trustworthy AI systems.