The forest industry gets an AI assistant built on expert‑verified knowledge
Digitalisation in the forest industry is accelerating, and data volumes are growing rapidly. A new AI assistant makes verified forestry expert knowledge easily accessible through natural language conversations.
The Finnish Forest Products Engineers’ Association has created ForestBioFacts, a global digital learning platform for the forest industry, which now features a new AI-based solution.
ForestBioFacts is a leading digital learning platform for the forest bioeconomy, with 1,700 articles across 16 themes and 50+ organisations worldwide using it.
The new ChatGPT-style learning tool for the forest industry introduces a conversation-based interface to the platform. It allows users to retrieve forest sector information directly through an AI assistant, making the knowledge base more interactive than ever.
“Information retrieval is no longer based solely on search terms, but on conversation. Users can ask questions, clarify details, and deepen their understanding just as they would with a human expert,” says Vera Hallikainen, Development Manager at the Finnish Forest Products Engineers’ Association.
Reliability of information is at the core of the new AI agent
The central premise of the solution is reliability. ForestBioFacts’ content is based on an extensive network of experts.
Approximately 200 experts from companies, research institutes, and universities in Finland and internationally have been involved in its development.
The knowledge base includes the Papermaking Science and Technology series, written by subfield specialists.
“The goal is to provide information that is quickly accessible yet reliable and contextualised,” Hallikainen emphasises. According to her, artificial intelligence also improves the discoverability of information.
“Users can ask questions in natural language, ranging from individual terms to broad topics, and at the same time build learning paths tailored to their needs.”
The forest industry’s own “ChatGPT”
Unlike open-web systems, the newly launched AI agent retrieves information exclusively from the ForestBioFacts learning environment.
“In the ForestBioFacts model, answers are based on curated content, and the user always sees which sources the information is based on,” Hallikainen explains.
“This ensures the reliability of the answers and the origin of the information. A registered user can ask questions about forest industry technologies, products, and processes and receive answers based on information produced by experts, which also includes links to the sources,” she notes.
The reliability of generative AI has been widely discussed in recent years, making this crucial.
Nature magazine, for one, has reported in multiple articles that large language models can produce plausible‑sounding but incorrect claims without source references. MIT Technology Review describes this phenomenon as “hallucinations,” in which AI fills in information gaps with its own guesses. MIT has called this a key obstacle to broader chatbot adoption.
A case in Finland in spring 2026 illustrated the consequences of such hallucinations. Two major news outlets published an erroneous drone‑related news story touching on national security. Both outlets stated that the mistake originated from incorrect information generated by an AI tool. In an opinion piece in Kaleva, M.D., PhD in Social Sciences and Adjunct Professor of Social Policy, Jorma Laitinen argued that the incident may reflect a broader systemic issue rather than isolated errors.
“AI agents are already such advanced representatives of their kind that, despite increased oversight and tool fixes, we can never be completely certain that they will operate solely in accordance with our objectives in all situations,” he writes.
Against this backdrop, ForestBioFacts’ forest industry‑specific AI assistant represents a different approach. Built on a closed, expert‑reviewed knowledge base, it generates responses only from curated content, with clear source citations for every answer. This increases transparency and directly addresses concerns about AI reliability.
Supporting skill development and access to knowledge
AI-assisted learning environments can significantly improve skill development, access to information, and decision-making. At the same time, however, the reliability of generative AI has sparked debate, particularly regarding solutions based on open online sources.
Hallikainen explains that ForestBioFacts’ AI assistant addresses this concern by basing its answers on expert-evaluated content and citing sources for every answer. Such solutions, based on a limited, quality-assured database, are still rare in Europe. Most generative AI systems use large, publicly available datasets that may include inaccurate or unverified information. This increases the risk of factual errors.
In practice, AI makes it easier to find information. A vast amount of material becomes accessible through conversation without a traditional search, and users can refine their questions as needed. This supports students, researchers, and professionals.
According to Hallikainen, AI helps both in searching for detailed technical information and in grasping broader contexts. It can also support the creation of personalised learning paths.
The solution reflects a broader trend in which AI is increasingly combined with a limited and reliable knowledge base. This approach is particularly important in fields where incorrect information can have significant consequences.
“AI is already being utilised in the forest industry to streamline processes and optimise production, but its role is also growing in data management, research, and innovation. ForestBioFacts’ AI agent represents a trend in which AI supports expert work and decision-making,” Hallikainen explains.
Going forward, AI is expected to strengthen continuous learning in particular. Learning will become more personalised, and information can be tailored to users’ needs, while expertise remains at the core.
For now, development of the forest industry-focused AI agent continues, Hallikainen says.
“Next, the knowledge base will be expanded to include historical publications and learning materials. In the fall, approximately 113,000 pages of new content will be added to the system, which will be searchable by the AI,” she says.