AI brings huge opportunities and new but manageable risks for the energy industry

Artificial intelligence (AI) and its subsets machine learning (ML) and deep learning (DL) are set to transform the way we manage and distribute energy so that power supply is reliable, affordable, and clean.

As with all technological revolutions, AI is creating opportunities and challenges for industries, public policymakers, and societies. 

Prominent among the opportunities is how AI has potential to speed the energy transition by helping to manage the increasing complexity involved in greater electrification and the grid integration of variable renewable energy sources (VRES), energy storage, and demand-side response. 

An intricate balance emerging between innovation and regulation, such as the EU AI Act, underscores the distinct opportunities and potential pitfalls for the energy industry. Being able to trust AI is the key to unlocking the opportunities and complying with existing and emerging regulation. 

High hopes for AI in the energy industry 

Companies throughout the energy value chain are responding to the opportunities. 

The benefits of advanced AI are being seen today in GE's enhanced wind turbine efficiency, Google's reduced data centre cooling power consumption, and National Grid's (UK) decreased damage to gas infrastructure. Grid optimization is illustrated by LineVision and Xcel Energy collaboratively deploying Dynamic Line Ratings, and by Stem's Athena software for managing virtual power plant (VPP) enhancing battery storage portfolios.  

Providers and suppliers of AI are hoping for explosive growth. Users/deployers are testing such products and developing their own. They are exploring if AI can boost competitiveness by improving efficiency and reliability, reducing costs. 

The European Network of Electricity Transmission System Operators (ENTSO-E) states that an AI approach can be implemented in ‘vast swathes of TSOs’ core business to adapt a large amount of data, providing support for decision-making either from a system operation perspective or within corporate development and business administration'. 

In ENTSO-E’s view, the main advantages and useful applications of AI for a TSO’s core business might be established in one of the following: using DL in drones for maintaining overhead lines; applying digital twins (virtual representations) of high-voltage equipment of high importance; introducing software in controlling and accounting to improve administration performance within an organization; and using optimization code for automated energy trading. 

Challenges to an AI revolution 

AI is witnessing a surge in applications within the energy industry. A recent DNV C-suite survey found that while only 12% of surveyed companies have already implemented advanced AI, 73% are actively piloting or planning to do so.  

Challenges to AI’s full and effective use in the sector include regulation impacting on AI use, ethical considerations, cybersecurity requirements, and technological and investment hurdles specific to the energy industry. 

Deployers/users such as transmission and distribution system operators (TSOs and DSOs) must also consider the risks they make be taking on in using AI from external suppliers. 

There is also a disconnect between the rapid pace at which AI is advancing and the understandably conservative pace at which new technologies are adopted by key energy industry actors such as TSOs and DSOs. 

For example, introducing a new Supervisory Control and Data Acquisition (SCADA) system for controlling, monitoring, and analysing a TSO’s assets and processes typically takes up to 10 years. Some legacy equipment in grids can be 30 to 80 years old. In addition, TSOs and DSOs may have products from many vendors, and these components and any systems in which they are found need to be quality assured and managed for risk reduction. 

Creating energy industry trust in AI 

Being able to trust AI is the key to unlocking the opportunities and meeting the challenges. 

Design and implementation of AI needs to be conducted by professionals who combine deep understanding of both the AI algorithms and the domain of application of the systems they work on. This expertise, which is in short supply and consequently a training challenge, is essential to mitigate risks associated with poor data quality and the outcomes derived from its analysis. 

Standards and recommended practices are also vital for risk management and to enable controlled technology adoption to allow industries to progress by building trust and efficiency. ENTSO-E, for instance, categorizes AI as having a Technology Readiness Level of 6 (i.e. demonstration stage) and says there are currently no ‘best practice’ examples from a TSO’s perspective. 

Assurance of AI-enabled systems 

The absence of best practice examples for AI use specific to the energy sector need not hold back innovation, however. DNV has invested extensively since 2021 to develop recommended practices (RPs) providing guidance for organizations and industries including energy throughout the digitalization journey. 

These RPs are guidelines and best practices for procurement, development, and operation of AI and other digital solutions. They include the recently published DNV-RP-0671 AI-enabled systems assurance, the world’s first RP on how to ensure the trustworthiness of AI-enabled industrial systems. 

The company’s roles and services related to digital trust include, among others, risk assessments; training and self-assessment; verification, validation and testing aimed at building trust in the applications. 

AI in the energy industry – the next five years 

If the challenges can be resolved, the rest of this decade can be pivotal for AI integration altering how energy is managed and distributed. 

Regulatory and trust issues may mean we see few high-risk applications of AI in the energy industry in the short term. However, DNV thinks it likely that the next two to five years will bring accelerated deployment of AI in smart grid monitoring, management, enhancing energy efficiency and reliability. AI-driven predictive maintenance in renewable energy installations such as wind and solar farms will become more prevalent, reducing downtime and optimizing energy production. 

AI applications in demand-response systems will become more sophisticated, enabling better balancing of energy supply and demand. We could see more AI in sensors and as part of monitoring equipment feeding data back to SCADA systems. 

Make haste, but slowly 

AI is poised to become a cornerstone of the utility sector as the technology’s rapid scalability offers an important tool for speeding the transition to cleaner energy systems. 

Beyond the next half-decade, we anticipate a paradigm shift towards fully autonomous energy systems, where AI not only monitors and manages energy flows but also makes strategic decisions to optimize performance and sustainability. 

In the power sector, the progression from analog to digitalization, then machine learning, and finally advanced AI is a continuous, incremental journey. It will certainly make some use of generative AI, the ‘public face’ of AI, for some technical and sales purposes, but is more focused on machine learning in an industrial operations context for the near future. 

In all futures, however, digital trust will remain the key to maximizing AI’s potential in an industry for which implementation is likely to be more a case of ‘make haste, but slowly’. It requires early involvement and a comprehensive systems approach covering both system-level and various digital building blocks, ensuring comprehensive guidance for the trustworthy, safe, and aligned integration of AI into operations.