Can artificial intelligence save real lives? Bjarte Johansen is working to find out.

Thousands of important reports are stored in Equinor’s databases, but if the right information doesn’t get to the right person at the right time, accidents can happen. Computer scientist Bjarte Johansen (31) is tackling the problem — using artificial intelligence. 

All photos: Ole Jørgen Bratland / Equinor

Bjarte Johansen is a computer scientist working for Equinor, at our Sandsli office in Bergen. At the age of 31, he’s written a doctoral thesis on automated analysis of Norwegian text, and he takes a passionate interest in how the Norwegian language is structured.

“Our language is very interesting,” he says. “For example, they don’t use the feminine class in Bergen, and they use different endings in ‘nynorsk’ (New Norwegian) compared with ‘bokmål’ (the more common written form of Norwegian). We also use different sentence structures and phrases depending on which part of Norway you are in.

“That’s why we meet some pretty unique challenges when trying to teach computers to understand what we write. So we can’t just describe the language using simple rules that the computer can apply; instead it has to learn how to interpret all the different nuances in our language.”

Bjarte Johansen

But what does this have to do with Equinor?
“We want to make sure our information reaches the people who really need it. For example, out on a platform the computer can suggest relevant incident reports to workers who are carrying out tasks and say why these reports are relevant.”

Developing Operational Planning Tool
Johansen is part of the development team working on Equinor’s Operational Planning Tool. On a quest for the best coffee in the building he has just escorted us around Equinor’s office at Sandsli in Bergen. After a summer of long hours spent at the office, he’s had plenty of time to locate the elusive elixir. He sips his piping hot coffee from a cup bearing the CERN logo, and explains:

“The computer works as a tool to assist users to make better decisions,” he says, and goes on to explain.

“A work order is prepared every time there is a job to be done out on the platforms, and this is given to the people who will do the actual work. Occasionally, things go wrong, and an incident report is written every time an abnormal situation occurs. The challenge with these incident reports is that they are stored in multiple systems, and even though we have routines to make sure that everyone takes note of relevant reports, there is no guarantee that this actually happens.”

Does Equinor know what Equinor knows?
“Now we have gathered all the reports in one system, and the system can say, “You really should read this report, because the last time someone did something similar, something went wrong,” he says.

While the underlying technology is advanced, Johansen himself describes it as relatively simple.

“It’s really all just about making computers understand the Norwegian language,” he says. He’d rather not claim that machine learning can save lives in Equinor — but he does concede that it can go a long way towards contributing to saving lives.

“Computers themselves don’t save lives. People save lives. But we can use computers to enable people to make the right decisions,” he says.

A computer that can detect hazards
Computers are quite exceptional at certain things humans are less good at, such as recognising structural patterns in huge amounts of information. And when we ask what people are better at than machines, Johansen answers that people have the ability to nuance the importance of information and semantics.

“No matter how well we teach the computer to understand the Norwegian language, there should always be a human making the final assessment of what should be done. That’s why the computer suggests relevant reports for the person who will perform the work, and it will give an explanation of why these particular reports are relevant. The computer functions as a tool the user can take advantage of to make better decisions,” he says.

What is machine learning?

Machine learning is the scientific study of algorithms and statistical models that computer systems use to perform a specific task without using explicit instructions, relying on patterns and inference instead. It is seen as a subset of artificial intelligence.

Enthusiastic platform manager
Bjørnar Skulstad is convinced that there are grounds for saying that the system that Bjarte Johansen and his team are developing, could save lives. He’s platform manager on the Oseberg platform, 63 years old, and a self-declared enthusiast of modern digital solutions.

So you believe that this expertise can be used to save lives in Equinor?”
“Yes. Without a doubt! Although that’s probably a slight exaggeration, there’s no doubt that this technology can contribute to greater safety on the platforms — and in that way it certainly can contribute to saving lives,” he says.

Bjørnar Skulstad

He’s full of praise for the new digital solution that Johansen and his team have developed.

“We have an incredible amount of experience and expertise in Equinor. The challenge has been that this experience and expertise has been mostly been reserved for the people who possess these skills. Now we’re finally organising all this experience into a system that will make it accessible for everyone. That is something we will all benefit from,” says Skulstad.

“Will help reduce injuries”
The average age in Equinor is high, many of the employees have developed a wealth of experience gained over many years when it comes to operational integrity and regularity in the company—and Skulstad has been very concerned about losing this expertise from the company. He says the new component in OPT will go a long way in resolving this concern.

“We have comprehensive training for all employees in the company, but after having worked here for more than 30 years, I can safely say that we humans never finish learning. Never. That’s why it’s so great that we can use digital solutions for even more learning, and that the machine can store, systematise and pass on this learning and the experience employees take with them when they no longer work here.

“In this way, all our precious expertise can be passed on to the next generation. This development is a giant step in the right direction, and in the future, computers will help reduce injuries and ensure that we have far fewer safety incidents,” says Skulstad.