DNV warns energy operators need trust frameworks before scaling AI
DNV has released a position paper arguing that energy network operators must establish stronger governance and assurance systems before deploying artificial intelligence more widely across critical infrastructure.
The paper, launched at the All Energy conference in Glasgow on May 14, addresses the operational risks that emerge as utilities increasingly rely on AI for managing complex, digitalized energy systems. Electrification, decentralized generation and renewable integration are pushing operators toward automated decision-making tools-but without proper oversight, these systems introduce new vulnerabilities.
DNV defines digital trust as confidence that AI systems will operate predictably, remain under control, and continue serving operational and public-interest goals throughout their lifecycle.
What operators are facing
Energy networks are becoming more dependent on AI and advanced analytics to handle growing system complexity. But concerns about model transparency, governance and confidence in AI-generated outputs are slowing adoption across the sector, according to DNV.
"AI is becoming integral to how energy networks run, but scaling it without assurance creates new and very real risks," said Hari Vamadevan, senior vice president and regional director for UK and Ireland energy systems at DNV. "Digital trust is what allows operators to deploy AI with confidence."
The risks span operational control, system resilience, data quality and regulatory oversight. Without stronger monitoring frameworks, broader AI deployment could undermine the stability that critical infrastructure demands.
Moving from pilots to operations
DNV pointed to its work on the UK government-backed Intelligent Gas Grid program as a practical example. The initiative uses digital technologies and AI to improve efficiency, safety and performance across gas distribution networks.
The company emphasized that AI assurance and governance must become central requirements for moving technologies from pilot projects into full operational deployment. That transition requires clear answers about how AI models work, who oversees them and what happens when they fail.
Graham Faiz, head of digital energy for UK and Ireland energy systems at DNV, said: "Ultimately, the energy transition will have to be digital, but it will only succeed if it is trusted."
For operations teams, this means building institutional knowledge around AI for Operations and understanding governance requirements before AI systems move into production. Teams working with energy infrastructure will need to grasp both the technical capabilities and the assurance frameworks that keep these systems accountable.
Your membership also unlocks:
