From Regulation to Implementation: EU AI Act Compliance in Engineering and Construction of the Oil & Gas and Petrochemical Sectors

Paper i proceeding, 2025

The growing integration of artificial intelligence (AI) within Engineering, Procurement, and Construction (EPC) operations in the oil & gas and petrochemical sectors faces new regulatory hurdles with the introduction of the EU AI Act. Many AI applications in these industries may fall under the high-risk classification due to their role in critical infrastructure, creating complex compliance challenges for practitioners. This paper examines the intersection between these emerging regulatory demands and existing industry practices, revealing significant implementation gaps. Through careful analysis of the legislation's technical requirements, we develop a practical framework tailored to the EPC industry's specific needs. Our approach distinguishes between three common AI implementation paths: commercial procurement, third-party generative model adaptation, and proprietary development. By restructuring conventional MLOps workflows to incorporate regulatory considerations at each stage, the framework offers actionable guidance for industrial practitioners navigating this evolving compliance landscape. This work responds to urgent industry concerns, as uncertainty about regulatory adherence increasingly hampers AI adoption in technically sophisticated sectors.