Artificial intelligence (AI) is often synonymous in the public lexicon with lost jobs and replaced humans.

But industry experts at Hart Energy’s Executive Oil Conference & Expo (EOC) say there’s no reason to worry. AI is simply here to be a tool.

AI is “not a replacement for what we do. We’re just trying to help increase productivity and decrease risks by giving people access to a lot of information in a way where it’s sort of pre-analyzed,” Jon Ludwig, president and co-founder of oil and gas data science company Novi, told the audience on Nov. 20 during EOC.

Ludwig and experts like him view AI as a valuable stepping stone to reach greater heights within the energy industry. Fellow panelist Sushma Bhan, director at Ikon Science, concurred.

“We can automate and augment human results, the skills SMEs [subject matter experts] bring, and bring that together with the automation. Things or actions that can be done in hours or days can be done in minutes,” Bhan said.

When looking at subsurface data, operators are confronted with hundreds or even thousands of pieces of information, whether it be images, documents or datasets. On top of that, subsurface data can be siloed, as many want to keep all information to themselves, Bhan said.

AI can take all that information, categorize it and simplify it, and then bring all of it together in a unified fashion and use the power of visualization to integrate in-stream data in real time.

“It’s about eliminating waste, it’s about expediting your decision making and it’s all about the bottom-line impact,” Bhan said. “We had one global customer where we had more than 86 wells, 650 images and we had to get that worked. And it was estimated that the work would take something like 13 weeks, but with the AI solution, it was done in one week.”

AI’s efficiency and impact to the bottom line is one of the biggest drivers for its increased adoption, but there are still some who are slow to come around to the idea of a machine accurately interpreting data. Many are so used to SMEs, company gurus and specialists that the validity of a machine’s findings are being questioned.

“I remember some of the earliest models we built [was] really good at telling people what the impact of putting more profit in the well is, but very bad at telling them what happens when you put wells closer together,” Ludwig said. “Building these models and looking at the results as a panacea is probably one of the biggest risks. You have to do what is necessary due diligence-wise to make sure that these things are putting out results that actually make sense.”

Part of due diligence is making sure that the data provided is high-quality data, with as much information and context as possible in order to create valuable insights, Bhan said.

“The crux of your success relies on your data’s readiness. What data are you using? Is it the right quality,” Bhan said. “Accessibility and quality of data is very key to make sure that the results coming from your algorithms and models are reliable.”

Other issues that are raised when it comes to AI implementation are issues with data security and data privacy–although those problems have to do with the operator more than anything–and the inherent biases that come with AI.

Since AI is based on and modeled after the human mind, they can come with the same biases that the person programming the machine had. These are things that need to be addressed and nipped in the bud before scaling up the technology, Bhan said.

Because of the aforementioned issues, operators implementing AI solutions need to manage expectations, Ludwig said.

“For any new technology, there’s always this initial exuberance that this is going to solve all of our problems,” he said. “Then you start trying to implement it pragmatically, and you find that you’re way over your skis, meaning you’re at risk of falling on your face constantly. And if you fall on your face, it can destroy a program. You can set it back a year or two.”

However, positives still outweigh the risks when it comes to AI, Bhan said. The technology is only learning and getting better.

“With AI embedded in our tools, especially if you have chat bots, they can quickly accelerate your efforts,” she said. “AI not only augments what the SME is thinking, but they also are learning. So if you have some decisions made and you want to summarize things, but you want your system to learn and be reused and redeployed, especially if you’re looking at scalability, that’s the easiest and the fastest way to go.”

AI is taking off at unprecedented speeds and has the potential to be a valuable tool for the energy industry. The world is changing, and the oil and gas industry needs to change along with it if it intends on keeping up, Bhan said.

“The biggest risk is not using AI at all. Embrace it.”

Comments