Set to separate winners from losers

CEO Karl Johnny Hersvik at Aker BP is convinced that the pace of implementing AI holds the key to future competitiveness, and that the biggest risk lies in not adopting it.

“Where we’re concerned, this has evolved from an interesting technological debate to being a ‘make or break’ issue for us,” the oil company head says. 

“We believe AI technology will come to create clear winners and losers in a way we’ve never seen before, not only in the oil and gas industry but also in other sectors.”  

Hersvik identifies two key factors driving AI development: its dramatic effect on productivity and the intense pace of change.  

“In principle, this technology is accessible to all,” he notes. “It’s the ability to apply it quickly and effectively which will separate winners from losers.”  

Direct 

Hersvik is known for his direct management style and burning interest in technology, which have both followed him from his maths studies and throughout his career in the oil industry.  

Nor is he among those who fear the trend.  

“AI doesn’t represent a threat, but rather something which offers us entirely new productivity gains – as long as we succeed in exploiting it in the right ways.”  

“It automates and greatly speeds up time-consuming tasks which were previously carried out manually,” Hersvik adds, and illustrates this point with a recent example.  

“A root cause analysis (RCA) which normally takes months can now be completed in just a few days. We tested this thoroughly with one we’d already done manually and knew the answer to. Running AI agents through the same analysis came up with exactly the same result at a time saving of 97 per cent.”  

In principle, this technology is accessible to all. It’s the ability to apply it quickly and effectively which will separate winners from losers.

Streamline 

This case shows how AI can streamline traditional work. But what do these dramatic gains mean for personnel in the organisation?  

According to Hersvik, people will become an increasingly important competitive factor. While machines handle repetitive work, data analyses and complex computations, employees can concentrate on activities which need human intuition and creativity.  

“I see this as getting an amazing new tool. The machine frees up time from routine tasks, allowing you to focus on strategic thinking, problem solving, innovation and intuitive assessments.  

“That’ll require new expertise profiles, where people must develop the ability to interact with technology and understand its opportunities and limitations.”  

Hersvik highlights three critical areas of expertise, including technology understanding – good insight into what is happening at the sharp end of development.  

“In addition come cross-disciplinary translation capabilities, which involve grasping both the technology and the relevant discipline domains.  

“The third requirement is management expertise – in other words, executives able to determine strategies, delegate resources and drive change forward.”  

In Aker BP’s experience, he says, the second of these is the most difficult to cultivate. It calls for people with a foot in both camps, who understand both the digital and professional domains.  

“The oil and gas industry will never become a leader in the actual development of advanced technologies such as language models. Billions of kroner are being poured into researching and developing these. Our strength must instead be the ability to adopt and apply these technologies effectively.”  

I’m incredibly enthusiastic about AI and I really don’t see any great risk in using it. The biggest risk lies in failing to do so.

Process 

AI is primarily used today for data extraction and decision support, rather than directly in managing production processes.  

“This is a matter of supporting productivity in basic process such as geophysical interpretation, reservoir simulation or RACs. AI can be applied across the entire industry,” says Hersvik.  

He also believes that the technology can be utilised with safety instrumented systems (SIS) – in connection with barrier monitoring, for example.  

Such solutions comprise three main components: sensors to measure condition status, logic solvers which interpret the data, and consequences or measures based on these interpretations.  

“Programmers traditionally use a fixed logic for each safety barrier, but an AI agent can handle thousands of such logics automatically,” Hersvik explains.  

He notes that many people are concerned about false positives – where an AI system mistakenly identifies a negative event as the opposite. But he believes that this risk can be significantly reduced.  

“False positives are a relevant concern if the system is trained using open and uncontrolled datasets. But utilising a high-quality, verified and limited dataset can go a long way towards minimising the problem. That makes it easier to develop reliable safety systems.”  

Hersvik also emphasises the importance of grasping that industrial data are fundamentally different, so experience cannot be transferred uncritically from general AI solutions to safety systems for industry.  

Biggest risk 

Hersvik highlights cyber security as the biggest risk when adopting new solutions.  

“Information leakage quite clearly poses a threat which needs to be handled,” he acknowledges. “Since cyber security is changing dramatically, it’s very important to have clear guidelines in place to control cyber risk and information leaks.  

“We must understand the technological barriers at all times, even though they’ll evolve extremely quickly. This doesn’t mean we can apply AI systems to everything. Industrial data and processes are too complex for that.  

“But I’d again emphasise that this isn’t really completely new. It’s a technology like all the others. The process, qualification systems and performance requirements are unchanged. What’s new is the pace of development.”  

Optimist 

While some are worried their jobs will disappear, Hersvik believes AI will become a new co-worker, collaborating with employees and boosting productivity.  

“That’s why I encourage people to be curious and to investigate this technology. We should respond to AI with a positive and forward-looking mindset.”  

Hersvik is also conscious of his own role as a chief executive.  

“I have a special responsibility both to drive technology development forward and to manage possible risks. Major upheavals must be led from the top. We can’t wait for this to develop organically from the bottom up.  

“So I’m devoting a disproportionate amount of my time to understanding the technology and becoming involved in test cases.  

“If I’m not on top of this progress, I won’t be in the forefront either of the biggest strategic revolution currently taking place in the world today.”  
He points out that companies and employees which successfully utilise AI in their processes will become far more efficient than those who fail to do so. That makes this a key strategic issue for managers in the time to come.  

“Be aware of what you’re doing, get involved as a leader, test the technology early and bring the organisation with you,” he says.  

“Because this isn’t primarily about technology. It’s about effective utilisation.

“I’m incredibly enthusiastic about AI and I really don’t see any great risk in using it. The biggest risk lies in failing to do so.” 

The oil and gas industry will never become a leader in the actual development of advanced technologies such as language models.  Our strength must instead be the ability to adopt and apply these technologies effectively.

More Dialogue articles on AI: