An ontological approach to foster the convergence, interoperability and operationalization of frameworks for Trustworthy AI

Fragmentation of AI safety frameworks is slowing progress: The paper identifies that while AI capabilities are advancing rapidly, AI safety research faces "critical challenges in terms of strategy, consensus and operationalisation," suggesting that the field lacks unified approaches to ensuring trustworthy AI systems.

A shared "language" for AI ethics could improve coordination: The AI-Ethics Ontology (AI-EO) uses semantic technologies to create a common framework that allows different trustworthy AI frameworks to work together and be understood consistently, addressing the interoperability problem that currently hampers the field.

Technical standardization can bridge the gap between principles and practice: By providing an "abstracted semantic infrastructure," the ontology aims to move AI ethics from theoretical frameworks to actual operational tools that organizations can implement, addressing a key gap between what experts recommend and what companies actually do.

The approach is designed for real-world adaptability: The ontology was developed through case studies and is intended to be "conceptually close to target applications" with built-in adaptability, recognizing that AI ethics frameworks must evolve as technology and society change.

Open access and accessibility matter for adoption: The paper makes version 1.0 of the ontology freely available and emphasizes usability, suggesting that responsible AI governance requires tools that are not locked behind paywalls or overly complex for practitioners to use.