As a new concept changes the approach to working with AI

How can artificial intelligence become a real partner in knowledge work - and not just a supplier of answers? This question is the focus of current research by Prof. (FH) Dipl.-Inform. Karsten Böhm, Professor of Business Informatics at the FH Kufstein Tirol. Together with Prof. Dr. Susanne Durst from Reykjavík University, he wrote the paper Collaborative GenAI - Humanized Interaction Fields for Knowledge Creation, in which they present their concept of Humanized Interaction Fields (HIF). The two then presented the results of their work at the renowned European Conference on Knowledge Management (ECKM).

The model describes a new form of interaction between humans and machines: a collaborative relationship in which both actors create knowledge together instead of just producing results at the push of a button. Böhm compares this collaboration with so-called cobots - collaborative robots that work hand in hand with humans in industry. In the digital space, humans and AI take on these roles: they share a common workspace based on natural language. “Our language becomes the connecting element of collaboration between humans and machines, building trust and understanding,” explains Böhm.

This change of perspective is key. While conventional AI systems are often seen as tools, the HIF concept places the machine at the center as an active co-creator. Generative AI models are therefore no longer viewed as pure computing machines, but as cognitive assistants that support the processes of idea generation, problem solving and knowledge development. The goal is a human-centered artificial intelligence that adapts to the needs, abilities and learning processes of its users - instead of replacing them.

KNOWLEDGE IS CREATED THROUGH COOPERATION

What is particularly exciting is that HIFs see knowledge work as a two-way learning process. While people gain new perspectives through the exchange with AI, the AI learns from the interactions at the same time. This creates a dynamic field in which knowledge is not simply generated, but co-evolved - i.e. developed further through interaction.

This means that humans benefit from the speed, structure and pattern recognition of the AI, while the AI can better assess how knowledge is contextualized and made usable for human learning processes through the feedback, formulations and decisions of the users. Böhm describes this dialog as “a kind of cycle of questions, answers, corrections and new impulses” in which both sides benefit from each other.

One example: When a person discusses a complex task with an AI system, the iterative exchange not only leads to a solution, but also to a deeper understanding of the issue. The AI reflects back the thoughts, structures them and offers alternative perspectives. At the same time, it learns through this interaction how humans think, evaluate and formulate depending on the context. In this way, a constantly evolving knowledge space is created - a Humanized Interaction Field.

BETWEEN SUPPORT AND SELF-LEARNING

However, Böhm also warns against relying too much on the supposed convenience of the machines. “As with training a muscle, AI must not do everything for us - otherwise we will forget how to think critically ourselves,” he emphasizes. He compares it to an exoskeleton: it can increase strength, but if it takes over the work permanently, the body develops muscle atrophy.

This also applies to the mind. HIFs should therefore support the human thought process, but not replace it. The goal is a balance: enough support to be productive - but not so much that the learning or cognitive process is stunted. Böhm calls this “positive cognitive reinforcement”, in which AI stimulates thinking instead of shortening it.

A NEW LOOK AT KNOWLEDGE MANAGEMENT

With this research, Böhm addresses central questions of modern knowledge management. His work builds on existing models - such as the SECI model by Nonaka and Takeuchi - which describes knowledge creation as an interplay between tacit and explicit knowledge. The HIF concept extends this approach by integrating AI as an equal actor in this cycle. The further development of the model is Generative Responsive Artificial Intelligence (GRAI), which is described in another publication.