HOUSTON – Generative artificial intelligence (Gen AI) is widely perceived and already being experienced by more mature markets as the next major ‘disruptor’ in energy markets – be it for predictive analytics for asset maintenance or optimizing power plant operations to empowering utilities on monitoring and managing real-time fluctuations in electricity supply and demand especially with massive integration of intermittent renewables as well as in having systemic energy trading algorithm.

During a discussion at CERAWeek, Accenture Chair and CEO Julie Sweet emphasized that more and more energy companies are now embracing generative AI as one of their top three priorities because they already realize that the sector is on to its seismic shift towards intelligence-driven innovation.

“There has been a massive mindset shift across industries from a couple of years ago to today as they understand that AI is absolutely critical component of their operation….I would say that similar to what we’ve seen in the digital transformation, we have some energy companies like Saudi Aramco, which use AI throughout their operations today,” she noted.

Typically, according to Gino Hernandez, head of Digital at ABB Energy Industries, the application of Gen AI is with individual assets of energy companies – and the key goals are optimization of their facilities’ performance and this is also a tool that firms are now using on their decarbonization strategies.

He explained that the correlation between AI to lowering carbon emissions lies in the fact that if individual assets of energy companies would gain more efficiencies in their operations through digital innovations – primarily on the AI space, then there will be less feedstock or raw material they will use in electricity generation, and that in turn will redound to carbon footprints reduction.

“The use of generative AI starts with the individual assets…we see a lot of customers who want to get better on optimizing their assets; and on the carbon emissions reduction perspective, it is also enticing,” he stressed.

Hernandez added that on the realm of asset operations and maintenance, “because of the flurry of information coming from the asset, it gives an opportunity to make business decisions that we don’t have before - trends like these are areas that are coming together for energy management and energy optimization.”

The ABB executive further stated “if you’re not thinking about generative AI,  how will you going to have your next competitive advantage factor? You’re going to be left behind, and that’s the situation you’ll be in.”

Grid of the future – with massive renewables

Patrick Raab, head of Business Development and Commercialization at Google X, similarly asserted that the incursion of Gen AI in the energy sector –purposively  for the envisioned ‘grid of the future’ has been coming “very exciting and it is very fast.”

He said “we are witnessing an expanding value chain that is underpinned by a grid transformation and if we look at this, we see the grid of today which is centralized, it’s analogue, it’s slow - for analogy, that is one-way highway.”

Conversely, he reckoned that “when we look into the future, moving to the grid of the future which is digital, it’s dynamic, it’s decentralized, then the analogy will be a multi-directional ‘super highway’…the grid that we are transforming today is the story of co-optimization with different layers of architecture and data chains.”

He specified with the wealth of data sets whipped up through Gen AI, “when we talk about planning and forward looking into the grid of the future – we understand the interconnection of different devices as well as sensors and multiple forms of power electronics across the supply side and demand side – it’s really the flexibility and the co-optimization of those different layers so we can be confident in our planning and studies to make those decisions – especially with the growing integration of renewables in electricity systems, we need to achieve that kind of flexibility in planning the grid of the future.”

Raab highlighted “the combination of these different components of the evolving architecture will allow interoperability to enable the understanding of data and to trust that data as well as to increase the granularity of the models and the understanding of how decision-making is made.”

He expounded “a key component of this is moving from commercial model predominantly deterministic to one that is probabilistic  -- and these different pieces and different parties in the grid infrastructure and protection of communication and data changes are necessary to be accurate and the granularity is significant and sufficient to make modeling analysis and decision-making for a future that is highly intermittent because of renewables - to really understand and see the flexibility of the future as valuable instead of it as a risk.”

The Google X executive underscored that for Gen AI to flourish as a solution in the energy sector, a trifecta of policy, innovation and investment must be fully accomplished across the value chain.

“Policy can play a role here as a catalyst participant as well as drive some of the common and open frameworks. And on the innovation side, it must be done across the entire ecosystem, from the raw materials to the end-users. And to achieve that, capital has to be deployed – the move from paternalistic to probabilistic grid, it really will be driven by being able to deploy capital and to innovate around commercial business models,” he noted.

Blockchain trading of electricity is likewise seen as another innovative domain in the energy sector where Gen AI is expected to be advancing by great proportions.

Rebecca Hoffman, president and CEO of Blockchain for Energy, enthused that while generative AI “is the scariest thing ever, it is also the most exciting – because in using it, things can get out of control, that’s why we need to have blockchain that will be tracking and tracing everything that you’re feeding into that model - ensuring that you understand what’s happening and then monitoring the performance and results coming out of that AI.”

Hernandez further conveyed that in ABB’s latest engagement with the hyperscalers, discussion revolved around having better codes and improvement in documentations when it comes to bot perspectives – fundamentally in comprehending context as well as inferring intent, adding that such “will require multi-faceted approach when it comes to natural language processing, machine learning and analysis to craft responses and actions that resonate through the company’s chain of operations.”