Oversight Methodologies for Artificial Intelligence Implementation in the Energy Sector

Abstract

Humankind has been intrigued by Artificial Intelligence (AI) for decades (Fenwick and Molnar, 2022) as it is foreseen as a tool to improve efficiencies, further organizational interests, and improve societal well-being. Over the past years, there have been advancements in developing and deploying AI systems and tools in many critical services sectors, influencing organizations and people with a mixed level of benefits, successes, and risks.
The research aimed to understand the current knowledge base and gaps concerning the compliance oversight for the safe development, deployment, implementation, and use of AI systems within critical infrastructure services, specifically focusing on the energy or electricity sector. The driving factor for this research is that within the critical infrastructure and services sectors, incorrect decisions influence more than financial returns but can cause severe equipment damage, premature failure, and harm to people.
The outcome of this research is a unique AI compliance audit framework development procedure, AI compliance audit framework, and AI audit process, which is a fit-for-purpose compliance solution driven from a senior executive management level, integrated into specific existing compliance or governance processes, which is more advanced than the existing AI governance frameworks, audit mechanisms, and checklists.
This research significantly contributes to the industry and sector by providing a practical structure that can be utilized to safely and sustainably implement AI systems in this critical sector. The AI compliance audit framework is a mechanism that allows the energy sector to place guardrails around the AI system that they are procuring or developing throughout its lifecycle. The framework considers a multi-dimensional audit regime, which can seamlessly integrate into existing quality, information technology, or environmental assurance processes. Notably, the framework ensures that the energy sector considers the integrated software systems collectively when auditing and ensuring compliance, not as individual components.
Lastly, the research provides a foundational structure for future researchers to expand on, focusing on gaps within the current governance structures, training regimes, and approaches to building a sustainable AI system environment.