Enabling Ethical AI: A case study in using Ontological Context for Justified Agentic AI Decisions

This is a full draft (preprint) of case study developed by Kaiasm as part of their participation in The Turing Way Practitioners Hub in 2024-25.

Abstract

Agentic AI systems, software agents with autonomy, decision-making ability, and adaptability, are increasingly used to execute complex tasks on behalf of organisations. Most such systems rely on Large Language Models (LLMs), whose broad semantic capabilities enable powerful language processing but lack explicit, institution-specific grounding. In enterprises, data rarely comes with an inspectable semantic layer, and constructing one typically requires labour-intensive “data archaeology”: cleaning, modelling, and curating knowledge into ontologies, taxonomies, and other formal structures. At the same time, explainability methods such as saliency maps expose an “interpretability gap”: they highlight what the model attends to but not why, leaving decision processes opaque. In this preprint, we present a case study, developed by Kaiasm and Avantra AI through their work with The Turing Way Practitioners Hub, a forum developed under the InnovateUK BridgeAI program. This study presents a collaborative human-AI approach to building an inspectable semantic layer for Agentic AI. AI agents first propose candidate knowledge structures from diverse data sources; domain experts then validate, correct, and extend these structures, with their feedback used to improve subsequent models. Authors show how this process captures tacit institutional knowledge, improves response quality and efficiency, and mitigates institutional amnesia. We argue for a shift from post-hoc explanation to justifiable Agentic AI, where decisions are grounded in explicit, inspectable evidence and reasoning accessible to both experts and non-specialists.

Acknowledgements

This case study is published under The Turing Way Practitioners Hub 2024-25 Cohort - case study series. The Turing Way Practitioners Hub works with experts from partnering organisations to promote data science best practices. In 2024, The Turing Way team welcomed Laim McGee and James Harvey as Experts in Residence to represent interests and opportunities for AI companies in improving business operations. We thank them for leading the development of this case study. Lucy Cull has conducted data analysis, and Andy Corbett is the technical reviewer alongside The Turing Way team members, Malvika Sharan and Arielle Bennett.

This work is supported by Innovate UK BridgeAI. The Practitioners Hub has also received funding and support from the Ecosystem Leadership Award under the EPSRC Grant EP/X03870X/1 & The Alan Turing Institute. The Turing Way Practitioners Hub’s 2024-25 Cohort was co-delivered by Dr Malvika Sharan, Senior Researcher - Open Research and Arielle Bennett, Senior Researcher - Open Source Practices. Lelle Demertzi is the Research Project Manager. The Turing Way Practitioners Hub, designed and launched in 2023 by Dr Sharan, aims to accelerate the adoption of best practices. Through a six-month cohort-based program, the Hub facilitates knowledge sharing, skill exchange, case study co-creation, and the adoption of open science practices. It also fosters a network of 'Experts in Residence' across partnering organisations. For any comments, questions or collaboration with The Turing Way, please email: turing-
way@turing.ac.uk.