The Impact of Generative AI
Trygve Nodeland and Jørgen Løken in Oulu, Finland. June 2026.
How can we map out and learn from the AI challenges and opportunities that wiki communities are facing? A sunny June weekend three WMNO representatives headed to the Wikimedia Northern Europe Meeting in Oulu, Finland, to learn and share.
By Elisabeth Carrera
We had been tasked with facilitating one of three workshops: What is the impact of generative AI on knowledge production in Wikimedia projects? This topic has been debated on Norwegian Bokmål Wikipedia for the last few years, and editors are compiling reading materials, discussions and case examples in an AI topic portal (in Norwegian). Recently the community decided to change the guidelines for article deletion, enabling admins to speedy delete articles that are AI generated and that have not been sufficiently edited and quality proofed by humans. Other changes to guidelines and workflows are being debated. We were eager to learn what’s being done on other language versions.
Wikipedia values under threat
At the meeting, Trygve Nodeland raised questions about the values under threat when AI generated texts are used in Wikipedia articles:
- How can you make sure that the text is not biased, and that all relevant views are represented?
- Sometimes AI doesn’t present relvant references, so how do you know that the sources are real and reliable?
- Is the article likely to be maintained and kept up-to-date if the sources are unknown?
- How can you guarantee that the conclusions are scientific, and not those of an AI?
- And what about hallucination?
Being a lawyer, Trygve Nodeland also commented on the risk of copyright infringement. AI generated texts and images will always be the result of copying other texts and images. So far, legally, this is a grey area.
Strategies for AI use
Elisabeth Carrera briefly presented some players and their strategic stands on AI use: The American Trump administration, Norwegian legislators, and the Wikimedia Foundation and their strategy on Artificial intelligence for editors. In Wikimedia projects, the biggest need for consideration and change is perhaps within community workflows:
Workshop session
The Norwegian trio, now completed with Jørgen Løken, was joined by other Wikipedians, and perspectives from four Wikipedia versions were shared. We discussed the effects of widespread use of AI tools, and perhaps most importantly: Which responsibilites can only be taken on by Wikipedia communities? Within our one-hour slot, we mapped out a few relevant aspects:
Different views
The workshop as well as casual conversations left us with an impression that most were in favour of regulating the use of AI on Wikipedia, and partially banning it. Some were more concerned with other pressing matters on Wikipedia, and some felt disempowered. Where would they start, if their community were to respond to the challenges of AI generated articles? A few were eager to make sure that future guidelines wouldn’t prevent them from using AI tools to efficiently collect information.
Clearly, the communities represented at the meeting are taking varying paths. Some are keen to evaluate and adapt their workflows and guidelines, whereas others trust that their current systems are sufficient. Hopefully, our mini workshop session left us all with a better understanding of AI challenges and opportunities within the Wikimedia projects.
Learn more
- The AI portal on Norwegian Bokmål Wikipedia (in Norwegian)
- Wikimedia Foundation strategy brief: Artificial intelligence for editors
- Wikimedia Northern Europe Meeting 2026
All photos: Elisabeth Carrera, CC BY-SA 4.0.
Category: News, Wikimedia Norge