By Lucca Elisa Pedersen Szast and Susanne Togeby, AAU Communication and Public Affairs
Image: AI-generated

AI has long been part of working life – so much so that we no longer think about the fact that it's artificial intelligence helping us filter out spam emails or controlling package-sorting robots in a factory. AI works particularly well in closed environments like factory halls. In contrast, we know very little about how AI functions in specialized professions such as medicine and architecture.

In the research project AI:Xpertise, researchers from Aalborg University are exploring how AI solutions can be developed and implemented within specialized professions—and what impact this has on the experts' work. The project focuses on both organizational and technological aspects and is jointly led by Kasper Trolle Elmholdt and Andreas Møgelmose, who are associate professors at the Department of Politics and Society and the Department of Architecture, Design, and Media Technology, respectively.

Why Researchers Want to Study AI Technology

According to the two researchers, a significant shift occurred when language models like ChatGPT emerged in 2022.

Kasper Trolle Elmholdt studies the use of artificial intelligence in organizations. He explains:

“AI gained real momentum with ChatGPT. Whereas we previously expected AI to take over manual tasks from factory workers or automatically send invoices, the idea now is that AI is increasingly expected to take on tasks we’ve traditionally seen as reserved for experts - or that require a close client relationship, whether it’s with a patient or an architecture client.”

But according to Andreas Møgelmose, who has worked with artificial intelligence for over 10 years, the actual changes brought about by the technology remain limited. He says:

“It’s deeply fascinating that machines can now understand and use language - something we once thought was uniquely human - and that has led to high expectations for the use of AI. But two and a half years after the breakthrough of language models, no one has really found the amazing business case - apart from a few niche areas like translation. Everyone talks about scaling AI across society, but no one has the recipe for how to do it successfully.”

Technology That Works in Practice

The goal of the research project is to develop AI tools that can actually be used in practice. It is well known that both technical and organizational factors influence the use of technology, but these aspects are rarely considered together. This is precisely the connection that AI:Xpertise aims to strengthen. The researchers seek to develop solutions that stem from real-world challenges faced by businesses and organizations.

For example, the team will develop an AI model to help architects visualize building projects for their clients, such as by generating 3D images of the house the architect is designing. Another example is using AI to assist doctors 
 
In addition to developing the technical solutions, the project also examines the organizational frameworks in which AI must operate. The researchers are closely following how the implementation is organized in order to understand what enables—or prevents—the technology from being used in practice. The focus is on workplace leadership and employees, but the researchers will also look at how patients and clients influence whether AI solutions are adopted and function in everyday settings.

The two researchers emphasize that there is no intention for AI to replace experts. Instead, the goal is to create tools that can support or take over certain parts of the experts’ work. The research aims to shed light on the new types of tasks and roles that emerge as AI is integrated into professional practice.

Translation by Emma Holtegaard Hansen