Wikimedia Futures Lab/Documentation/Prague 2026

Documentation: Wikimedia Futures Lab, 26.4, 2026, Prague

This is the documentation of the Wikimedia Futures Lab workshop that took place on April 26, 2026 in Prague. You can find the slides here.

Experts Inputs

Opening Keynote

Keynote Josef Holý

Josef Holý, Lecturer on AI and co-host of Canaries in the Ned, a podcast about algorithmic propaganda

“I would like to urge you, stop thinking about Wikipedia as a website. It's not a website, it's a cognitive power center".

In his keynote, Josef presented Wikipedia not just a platform competing for attention, but a model for democratic knowledge infrastructure in an era where truth, power, and technology are deeply intertwined.

AI Panel

AI panel

Across Europe, various labs are building AI systems that offer alternatives to the biggest commercial AI models. What could they look like, and what can they learn from Wikipedia?

Jan Hajic, co-founder of the OpenEuroLLM project, the largest EU consortium dedicated to developing open LLMs
Jan Šedivý, head of Big Data and Cloud Computing at CIIRC (Czech Institute of Cybernetics and Robotics)
Marian Grubben, former EU Commission staff member and current board member of Wikimedia Europe

Moderated by Alek Tarkowski, Director of Strategy at Open Future Foundation

This panel explored how AI is reshaping knowledge creation, access, and governance. Panelists emphasized that while AI can enhance how knowledge is produced and consumed, it also shifts the core challenge toward verification and trust. Open, community-driven models, like Wikipedia and emerging open-source AI initiatives, were framed as crucial alternatives to dominant, centralized systems.

Lightning talks

You can find the slides here.

Readers and Contributors Data in Europe, Klara Joklova
Approach to innovation at Wikimedia UK, Daria Cybulska

Readers Panel

Readers panel

How can online knowledge platforms remain visible, relevant and sustainable as AI reshapes how people access information online, and what are the implications if they fail to do so?

Ali Mahmood, service Lead for Audience Development at the International Center for Journalists in Prague
Katarzyna Drożdżal, UX designer and founder of Selkie, a Polish research agency for digital products
Tereza Schambergerová, Head of SEO at Seznam.cz, the largest Czech internet portal and search engine
Viet Hung Nguyen, Commercial Director of Online Development for FTV Prima, the first private TV channel group in the Czech Republic

Moderated by Lucy Crompton-Reid, Chief Executive of Wikimedia UK

This panel focused on the decline in human readership of Wikipedia and how shifting user behavior is reshaping how people access knowledge. The discussion emphasized that maintaining relevance will require adapting to changing user behaviors, strengthening relationships with audiences, and evolving organizational practices, while staying true to Wikipedia’s core values.

Hypotheses Creation Workshop

Sense making in small groups
- in 4s discuss: Of all the trends presented so far, which ones do you see in your Wikimedia work? What impact are they already having, and how might this change in the future?
Generate first hypotheses in small groups

Silent work on Canvas
Discuss in groups of 4-6 ppl: What's changing: What trend, challenge or opportunity are you focused on? What impact is it having? How should your community or affiliate respond to it?
Bring 1 hypothese per group to the plenary

Facilitated conversation around participant's hypotheses in the plenary:
- New interface: "Users prefer AI not because of AI but because of the UX experience and interface. People want conversational access to knowledge".
- Reader UX on Wikipedia: "Users are overwhelmed with current digital landscape, Wikipedia should provide smaller pieces of information like FAQ of articles and summaries of articles"
- AI policies on Wikipedia: "AI will amplify existing bias against minorities in the Wikimedia movement, and we need to oppose that by promoting Wikimedia projects as AI-free-zone (GenAI and algorithmic suggestions, tools could be admissible)"
- Brand visibility: "People are not aware of the uniqueness of Wikipedia. We need to change the narrative and create one that is accessible to a broader audience."
- AI tools for moderation: "We can use AI tools to detect harassment and behavior issues on talk pages. LLM are great to understand language, let’s utilize this".

Experiment Barcamp

Two topics were discussed:

Format for community workshop on AI policies on German Wikipedia - Alice Wiegand and Nicole Ebber
Lab in a box - Jan-Bart de Vreede and Eva Martin