Ethical AI, RIM and the Role of Ontologies
DOI:
https://doi.org/10.26686/arch.11630Keywords:
AI, Ontologies, Records and Information ManagementAbstract
This article discusses how ethical AI governance can be supported by records and information management (RIM) processes and practices, particularly the use of relational ontologies. The role of RIM in ethical AI governance is, and will increasingly be, vital in ensuring AI content is compliant with legislation, and that it demonstrates the characteristics of reliability, integrity, authenticity and usability as set out in RIM standards such the ISO 15489 Standard for Records Management.
This narrative is deliberately general and aspirational. It aims to provide an opportunity for further thinking and debate around the ethical use of AI, and the vital role that records and information management plays in its governance. To do this, it summarises content from some of my previous articles on the topic of AI, RIM and ontology, it also introduces some of my more recent thinking on the topic. It starts by covering what ethical AI is, and the role of RIM in achieving these aims. It then moves onto a summary of the role of ontologies in RIM, why they can be the ultimate underpinning design model for AI, and why they are essential in RIM as we move from traditional manual paper paradigm practice towards digital-native, automated modes of operation. It then concludes with some brief discussion of ontology based RIM approaches for ethical AI.