From “Should we?” to “How do we?”: A community weekend on AI and Wikipedia

When artificial intelligence comes up on Wikipedia, discussions can quickly become polarized. Should AI-generated content be allowed? Does it threaten quality? And can volunteers even spot it? Over the past two years, questions like these have shaped countless debates in the German Wikipedia. In February 2026, the community wrapped up a Request for Comments and agreed on a restrictive framework for AI-generated content. The vote clarified the rules—but it also made clear that many practical questions were still open. So instead of organizing yet another principles debate, Wikimedia Deutschland invited the community to spend a weekend tackling those practical questions together. Creating space—without steering the discussion Our aim wasn’t to influence the community’s position on AI. It was to create the conditions for a constructive, productive conversation. Ahead of the event, participants suggested topics through a survey and received background material to help them prepare. This ensured that participants started from a shared foundation and helped address one of the challenges identified beforehand: varying levels of knowledge about AI within the community. During the weekend, Wikimedia Deutschland provided structure, facilitation, and technical expertise—but not the answers. Our role was to bring people together, connect perspectives, and enable the work. Around 40 experienced volunteers met in Berlin for a structured co-creation workshop. Instead of long presentations, the weekend combined short inputs with self-organized working groups, open discussion formats, and collaborative documentation. Two facilitators helped keep conversations constructive, while leaving participants in charge of shaping the results. To move beyond the familiar “pro AI” versus “anti AI” framing, the workshop was organized around four complementary perspectives: - Wikipedia without AI – quality assurance, moderation, and policy implementation - Wikipedia with AI – legitimate use cases, tools, and workflows - Wikipedia on AI – improving encyclopedia content about AI itself - Wikipedia despite AI – preparing for long-term developments for Wikipedia and its communities From policy to practice One of the biggest surprises was how much the conversation had shifted. Only a few months after a highly emotional Request for Comments, participants spent very little time revisiting whether AI should be allowed at all. Instead, they focused on what implementation could actually look like: - How can AI-generated content be detected? - Which AI-supported tools could strengthen quality assurance? - What kinds of AI use fit with Wikipedia’s principles? The meeting produced a wide range of concrete proposals—ideas for review and monitoring tools, AI-assisted quality assurance, structured maintenance workflows, dedicated workspaces, and future research questions. Participants also suggested creating permanent spaces where AI-related issues can continue to be discussed and developed collaboratively. One participant summed up their experience afterwards: “As someone who generally supports the responsible use of AI, the meeting helped me better understand perspectives that are critical of AI. It also helped me clarify where I personally draw the line between appropriate and inappropriate uses of AI in an encyclopedic context.” For us, that quote captures one of the weekend’s most valuable outcomes. Success wasn’t about reaching agreement—it was about creating a space where people with different perspectives could better understand each other, while working together on shared challenges. Another important outcome was that the conversation broadened. Discussions about AI in the German Wikipedia had often been driven by a relatively small group of highly engaged contributors. The workshop created an opportunity for many more community members to join the conversation, contribute their own perspectives, and build a shared understanding of the challenges ahead. Looking ahead The discussions also showed that many community members are ready to engage with these questions in a nuanced way: rejecting unchecked AI-generated content while exploring carefully controlled forms of AI-assisted work. This pragmatic shift suggests that future progress will depend not only on policy decisions, but also on continued learning, shared processes, tools, and spaces for collaboration. For the Wikimedia movement, the meeting offers a useful example of how communities can approach AI-related challenges through structured participation, open exchange, and a focus on actionable outcomes. As AI becomes an increasingly important governance issue across Wikimedia projects, affiliates can play a valuable role—not by deciding community positions, but by creating the conditions that help communities develop solutions together. The project page and documentation are in German only. For further questions don’t hesitate to reach out to Verena. Can you help us translate this article? In order for this article to reach as many people as possible we would like your help. Can you translate this article to get the message out? Start translation