AI must be learnt collaboratively

When SDU IT offers courses in artificial intelligence, places fill up quickly. The high demand has meant that the AI Hub has primarily focused its efforts on developing and running courses. Consequently, work on communities of practice for AI has started later than planned.

The communities of practice form part of SDU’s decision-making framework for AI, which the executive board approved in September 2025. They are intended to bring together staff with shared professional interests or challenges and give them the opportunity to explore how AI can be used meaningfully and responsibly in their work.

– We have had to prioritise meeting a very specific need for courses. Every time we’ve offered a course, there’s been a waiting list. Now we also need to create spaces where staff can build on their knowledge and experience together with others, says Hans Lind, community builder at the AI Hub.

More than just a professional network

A community of practice differs from both a course and a traditional professional network.

On a course, the content and learning objectives are usually set in advance. In a professional network, participants typically meet to share the knowledge they each already possess. The community of practice lies somewhere between these two forms.

– In a community of practice, we share the knowledge we have, but we also create new knowledge together. We learn from each other’s experiences and come together to develop a practice, says Hans Lind.

Participants can, for example, show each other how they use AI in specific tasks, provide feedback on working methods or explore a shared challenge. A community can also work together to develop a tool or an agent for a specific work process.

The format is not fixed. Some communities may have a specific purpose and meet frequently over a shorter period. Others may continue over a longer period with longer intervals between meetings. The key is that participants come together around a shared professional interest and contribute actively.

The format is still evolving

There are already a few local communities of practice and pilot schemes at SDU. The AI Hub continues to explore which themes and working methods are relevant, so that staff prioritise taking part.

Possible themes include AI in administration, research support and curriculum development. The AI Hub initiates some of the communities, but they do not necessarily have to be run by SDU IT. Hans Lind sees particular value in communities that arise from specific needs within local academic environments.

– We can propose a framework and help get a community off the ground. But it is the participants’ interest and commitment that determine whether it takes off. A community of practice cannot be established solely by a central decision, he says.

Communities of practice are not intended to replace courses. The courses provide staff with a basic understanding and a common language. Within the communities of practice, they can build on what they have learnt and link it to their own tasks and professional standards.

AI use must be brought to the fore

The use of AI raises professional questions which staff would benefit from exploring together. Does the technology free up time, or does it simply shift the workload to checking and corrections? Does the technology support professional practice, or does it become an inappropriate shortcut?

According to Hans Lind, it is crucial that such questions are discussed openly.

– The worst thing that could happen here at SDU is for each of us to sit and work with AI on our own, behind closed doors. The best thing is for us to tell each other how we’re using the technology, look over each other’s shoulders and talk about when it helps us – and when we perhaps shouldn’t use it, he says.

Transparency is one of the key principles of SDU’s AI code of practice. The communities of practice can help translate this principle into professional discussions of specific experiences and working methods.

At the same time, they can make the knowledge that already exists at the university accessible to more people. Some of this knowledge is closely linked to individual staff members’ experiences and is not necessarily shared through the usual channels.

– There is already a great deal of knowledge amongst staff members. When we share it and learn from each other’s experiences, it benefits more people. AI is a team sport, emphasises Hans Lind.

Staff can take the initiative themselves

Throughout the autumn and into 2027, the AI Hub will continue to work with several communities of practice. One of the planned initiatives is a Copilot club, which you can sign up for now.

The AI Hub is also regularly producing guides for staff who wish to set up their own community of practice. But there is no need to wait for the guides.

If a small group shares an interest or is facing a common challenge, it can contact the AI Hub right now and get help finding a format that suits the group’s needs.

Read here about how to set up your own AI community