Forestry using AI

I’ve been around long enough to remember a time without the internet, and when mobile phones were just about portable. Some people had a computer at work, but internal mail consisted of brown circulation envelopes, not emails. Nevertheless, people began to voice ever-greater hopes for ‘the new technology’.

It was widely accepted as a useful term. It certainly referred to technology, and it was definitely new. The term was also useful because it could be applied to everything, and didn’t need to be specified in any great detail. The term probably lasted a couple of decades before it reached retirement age. In the meantime, a slightly more specific term had emerged, namely IT. That term promised universal salvation, and whatever challenge was raised, the solution was always ‘we can do that in IT’.

This puzzled me for a long time, and it still does today, there being two entirely different reasons for this. On the one hand it is fantastic that so many people, for such a long time, can place so much faith in a concept that continually refers to something vague that is coming soon, but not now. On the other hand, it is unsettling to say that you are going to do something ‘in IT’ rather than ‘with IT’. I can’t possibly be the only person on earth who has ever had these two thoughts.

Nowadays, everyone has at least one computer and mobile phone, everything is in the cloud, and customer service has been replaced by a chatbot. Still, vague hopes for the future remain. The only difference being that the concept has changed its name, and the hopes are much greater. Today, it’s all about ‘digitalisation and AI’. But fortunately, not ‘in’.

It’s curious how the terms ‘digitalisation’ and ‘AI’ go hand in hand these days. Digitalisation has been around for a long time, it has a fairly clear meaning, and concrete tools are available for both individuals and businesses. Some are said to be business-critical. But AI is quite new and still a bit fuzzy. Yesterday, AI was a strange word that perhaps meant neural networks or machine learning. The application was called Deep Blue and it could beat a human chess grandmaster. Today, AI is a buzzword that perhaps means large language models, and the application is called ChatGPT and can beat just about any human being at information management (and can hallucinate a bit at times)

But what will AI be tomorrow? It is clearly a phenomenal capability, but what is it for? What specific AI tools will companies use? Which ones will be business-critical? And what consequences will this have for companies, their employees, their customers and for society?

The digitalisation of forestry has been underway for some time. Sweden can claim to be a leader in this field. There is a range of digital tools currently in use by companies, and even more that are at various stages of research and development. When these are introduced, we will gain slightly increased efficiency and somewhat altered work processes.

But now, even the forestry sector is talking about “digitalisation and AI”, with the AI aspect accounting for an ever-increasing share of the optimism, even as its contours remain rather vague. Development is rapid and exponential, and enormous hopes may well be realised. Progress may be slow at first, but will accelerate after a while. However, this time it might not stop at new technology and somewhat altered work processes.

Research into forestry must incorporate everything that AI can offer, which is likely to mean huge technological leaps and far greater changes to work processes, to business models and employees, to customers and to society.

Fortunately, I haven’t heard anyone say “we can do that in AI”, but it’s only the linguistic error I’ve been spared. There still remains a vague hope for something undefined in the future. It’s time to lift the fog surrounding AI, and produce concrete pictures of what it might mean when AI makes a fully-fledged breakthrough in the forestry industry.

Per Edström

Deputy Programme Director, Mistra Digital Forest