How could AI help (and hurt) forestry?

(Press-News.org) The whole world is buzzing about the potential and pitfalls of artificial intelligence—including those who work in forestry.

AI could revolutionize forestry, making it possible to save more lives and ecosystems through faster and more accurate data analysis. But if forestry professionals aren’t careful, AI could also botch critical land-management and policy decisions. 

That’s why NAU School of Forestry faculty members Alark Saxena, Luke Ritter and Derek Uhey took it upon themselves to understand foresters’ relationship with AI: how they’re using it now, how they hope to leverage it in the future and what concerns them. They conducted 20 in-depth interviews with forestry professionals in the Southwest and recently published their findings in Forest Policy and Economics. 

“We noticed a great deal of discussion about the potential of AI in forestry, but very little research on how the professionals on the ground actually feel about it,” Saxena, an associate professor of human dimensions of forestry, said. “This study was our first investigation into the topic, motivated by the need to understand the human side of this technological shift.” 

In their interviews with foresters working across academia, government and private industry, the research team discovered that no one in forestry wants AI to replace human expertise or make critical decisions without oversight from real people.

“They are particularly concerned about the ‘black box’ problem where they can’t understand AI’s decision-making process, creating serious accountability issues,” Saxena said. “A key concern they shared was the risk of training AI using some agencies’ poor-quality or biased data and then trusting its flawed outputs for important land management or policy analysis”—like mandates on where to administer prescribed burns or allow clear-cutting.

But the forestry professionals they interviewed agreed AI could be a useful tool in supporting some aspects of their work. With current labor shortages leaving them overworked and burned out, the workers agreed they’d welcome AI help with monotonous tasks like summarizing information, lesson planning and filling out routine paperwork.

“They also see great potential in using AI for complex data analysis, such as with light detection and ranging, as long as it functions as an assistant that enhances, rather than replaces, the judgment of an experienced professional,” Saxena said.

Ritter said he hopes others across the United States and the globe will conduct interviews with forestry professionals to capture a wider range of perspectives on AI. Getting a full understanding of professionals’ fears about—and recommendations for—the use of AI could help leaders create policies that guide future forestry work.

"It's challenging to ethically implement AI when we, as foresters, have gaps in our knowledge about how and why it's being used,” Ritter said. “This study highlighted some interesting themes, but we need to keep discussing AI in the classroom and the workplace. We hope this paper provides a foundation for policy changes and further research as AI continues to grow."

END