Artificial intelligence (AI) technology offers massive opportunities to the oil and gas industry, and companies are identifying how to effectively deploy AI projects.
While exploring the themes of AI and human roles, trust, and potential during CERAWeek by S&P Global in Houston in March, industry experts discussed the spectrum of how companies view AI, how the technology could redefine how the industry works, and what it takes for an AI effort to be successful.
Speaking during the “AI in Energy: Managing the Transformation” session on 23 March, Conner Prochaska, director of the Advanced Research Projects Agency-Energy at the US Department of Energy, said AI is transformational and “everybody is using it and interacting with it in a real way right now. I don’t know if there’s a switch that happened. I mean, people talk about the ChatGPT moment, etc., etc., but the fact is, it’s here. We’re doing it.”
Gino Hernandez, ABB’s head of digital business for energy industries, said where a company is on its autonomous journey is a factor in how much it benefits from using AI. Many start with a focus on asset performance before moving into augmenting operations and then on to closed loop and agentic AI applications, he said.
Hector Rocha, global generative AI leader for energy at Accenture, said many companies are “looking for that magic wand” when they think about deploying AI technology.
“Many companies have started with lots of pilots, lots of POCs (proofs of concept), trying to unlock, trying to feel the AI, trying to test it, etc.,” he said.
What differentiates the companies leading in AI from the rest, he said, is the leaders know what business issue they are trying to solve.
“Not everything has to be AI,” he added.
That raises the question of what should be handled by AI, Prochaska said.
“Where is the value? Where is the work that needs to be done by AI versus over-qualified humans?” he said.
Human or AI?
The industry is witnessing a shift in trust when it comes to AI, Intuigence AI CEO Mo Tanabian said during the “Role of Humans in an AI-First World” session on 24 March.
“The observation I think at this point, as our customers are deploying and getting used to how to use it as a tool, is, ‘Oh, I can actually trust AI,'” he said.
Vivek Chidambaram, Accenture’s global resources strategy lead, said that, while AI is gaining trust, a human is still most often in the lead.
“We think of it as human in the lead, not human in the loop,” he said.
Gin Kinney, executive vice president and chief administrative officer for NRG Energy, said keeping humans involved in AI decisions is critical in certain circumstances, such as if it could cause irreversible harm.
“AI can deliver the insights, but a human has to retain the authority to make those decisions” using critical reasoning and sound judgment, she said.
Accountability is critical, she said.
“Someone needs to be held accountable. I don’t know if we’re ever going to depose Claude or call ChatGPT to the witness stand,” she added.
At the same time, she said, personnel may become more anxious about making decisions.
“Some of that inherent fear is also getting them like, ‘Oh, now the decision is up to me. Have I learned the right critical thinking skills in the course of my career, or have I just been doing tasks?’” she said.
And handling menial, unwanted tasks is one of the sweet spots for AI.
Chidambaram said AI could seem indispensable to someone if it handles that 20% of the job they hate doing.
Where Is Work Done?
While many companies have had disappointing results with AI efforts, the technology has delivered significant financial benefit to ExxonMobil in the past 5 years of systematic deployments, Rob Crane, technology scouting and venturing manager at ExxonMobil, said during the “Democratization of AI: Redefining Where Work Gets Done” session on 23 March.
“In the past 5 years, we’ve deployed it across quite a number of applications for a total value of over $500 million that we’ve delivered through AI,” he said. “By having that intentional systematic approach, you can get much higher percentage hits on it than the MIT study that talks about the 95% failure” rate of AI deployments.
AI can elevate the capabilities of people within the organization, he said, but it’s important to consider the informational, reputational, and operational risks of AI use.
“We want things that can make our lives easier. We want things that can help us be more productive. You always have this sense of this fear of missing out if others are going faster. But what we’ve learned, especially from the digital journey, is that, taking a systematic approach, these things can really pay off as opposed to adopting in an unsystematized way,” he said.
