Energy and utility companies are embracing — or at least considering embracing — generative artificial intelligence, per two recent surveys of the industry.
- The top line: According to a survey from tech giant IBM, 74% of energy and utility companies have implemented or explored incorporating generative AI into their work. But the road from interest to deployment is a long one, and there are several market trends at play that are slowing down adoption.
- The market grounding: While utility companies tend to be reticent to adopt new technologies, 2023 has seen a growing number experimenting with its use for everything from modeling future load to improving smart meter technology. And the energy sector has also embraced its promise in areas like automating microgrid operations and detecting battery defects. The use of generative AI, though, has increasingly appeared in comparatively low-risk contexts, like customer service.
- The current take: When it comes to implementing new tech, utilities “love to be a fast follower,” said Casey Werth, IBM’s GM for the energy industry. “There’s a huge expectation, and there’s a lot of experimentation,” he added. But for that excitement to make a difference, new AI capabilities need to be truly integrated into utility processes in the relatively short term, “and that comes from proving out the trustworthiness and the governance around these tools.”
Of the energy sector respondents who said they were interested in generative AI, just 27% are actively implementing it, the survey found, while 47% said they were “exploring” use cases.
One key reason behind that gap is safety concerns, Werth said. A separate IBM survey of CEOs in the energy industry found that 61% are concerned about the sources of data used in generative AI in particular.
Werth explained that many utilities are restricting their experimentation with generative AI to projects that can be completed with publicly available data. Utilities are loath to “expose” their data to SAAS, he added, and there’s “definitely some concern” about the level of data access required to train certain tools.
Another potential adoption barrier is utility desire to stick with existing vendors rather than building new products, Werth said, even if that means waiting for a product that hasn’t yet been released.
“I think this expectation of potentially being able to inherit capability from vendors is slowing some CIOs from sprinting ahead and building new,” he said, adding that that’s not necessarily a bad approach: “why build when you can inherit it?”
Ultimately, one of the greatest “stumbling blocks” utilities have experienced in the last six months when it comes to implementing generative AI is their lack of an amenable governance structure, Werth concluded.
Utility-wide implementation of generative AI capabilities, whether for resiliency planning or outage management, will require utilities to become data-first companies, and on that front, there’s a long way to go.
“We will stay nascent until we start to drive that philosophical change, or that change of identity,” Werth said.
IBM’s Global AI Adoption Index survey interviewed over 2,300 IT professionals from companies across 20 countries. And the IBM Institute for Business Value’s 2023 study titled “CEO decision-making in the age of AI” interviewed 420 CEOs of energy and resource companies. While both surveys were published previously, IBM unveiled the energy-specific findings in conjunction with this week’s Distributech conference.
Latitude Intelligence is soon to publish its first report on the use of AI by utilities. This joint research program with Indigo Advisory Group is a first-of-its-kind study of the pathways to adoption of AI-based solutions in the power sector. Through multiple interviews with utilities across the US, from investor-owned utilities to public power, this research uncovers how deployment strategies, existing applications, and targeted benefits are evolving. Sign up here to be notified when the report is released.
