Robots Thrive in the Forest on Jobs That Humans Find Too Boring

From watching pulp cook for hours on end and tracking parasite bugs on satellite photos to handling lengthy legal documents, Swedish forest companies are creating new jobs they would never ask a human to do.

Packaging maker BillerudKorsnas has been an early adopter of artificial intelligence by using the technology to analyse thousands of diagrams to determine just how long it needs to cook its wood chips before they turn into pulp. While that process could be done manually, it says it would be difficult to find any human who'd be willing to spend all day just looking at such charts.

"A machine can review large data quantities and find patterns in ways we humans just find too boring," Olle Steffner, director of intellectual property management, said. "Tasks such as monitoring processes or analysing diagrams will hardly be missed by anybody. Our staff is needed for other things."

The rewards from using AI for such mundane tasks could be plentiful. The biggest advantages include being able to replace costly manual labour with automation as well as reducing the time machines used in the manufacturing process are idle for maintenance, said Joakim Wahlqvist, who develops AI solutions at consultancy firm Sogeti. Companies can also use AI to help them improve the manufacturing process, as BillerudKorsnas is doing, he said.

Sweden's forest companies are the latest example of an industry embracing artificial intelligence to cut costs and lift profits. The country's banks are developing robotic customer advice and services such as chatbots and fashion giant Hennes & Mauritz is using AI and big data to foresee trends and optimise its logistics chain. While the forest industry still lags retail and manufacturing in using AI, it still has the strongest drive to automate among the traditional process companies, Wahlqvist said.

Sogeti has together with Sveaskog, Sweden's largest forest owner, developed algorithms that teach themselves to find signs of spruce bark beetle attacks on satellite photos of forests. With the bug threatening to destroy wood valued at SEK 6 billion ($622 million) in a worst-case scenario in Sweden this year, AI could become one of the most efficient defences against the bug.

"You could gain the same knowledge by putting on a pair of boots and walking into the forest to check for yourself, but AI helps you to attain it without the cost for large amounts of manual labour," Wahlqvist said.

Stora Enso, one of Europe's largest paper and packaging makers, has found another application for AI by taking pity over its legal department. It's trying to teach an algorithm to identify risk in the vast amounts of contracts it is handling, which would free up lawyers' time. It has already concluded its first trial comparing AI results with lawyers' assessments and mulls taking the work further with a new project. It's also using AI to analyse pulp at a Finnish mill.

A recent development with both cheaper computer-processing power and more advanced sensors is what is making the technology more accessible to the forest industry right now, according to Steffner at BillerudKorsnas. The company is stepping up efforts with a two-year long program that will look at all its processes and evaluate where they can be made more efficient by using robots.

"Where problems demand AI to be solved, we now have the knowledge to do so," Steffner said.

© Thomson Reuters 2019