Wikipedia has been around for almost 25 years. During this time, the site has become the most important free source of knowledge on the internet – with over 65 million articles in 300 languages, edited by hundreds of thousands of volunteers worldwide. But since the advent of artificial intelligence, a lot has changed. At this year’s WikiCon, the German-speaking communities addressed current challenges such as declining visitor numbers and AI-generated texts.
More and more people are using generative AI programs such as ChatGPT to search for answers. This has also created new challenges for Wikipedia: on the one hand, fewer people tend to access the site directly. This is especially true for the younger generation. On the other hand, large AI language models are increasingly drawing on the knowledge in the online encyclopedia. Without free data, these programs cannot be trained.
This shows that Wikipedia is and remains relevant for anyone seeking knowledge. But access to it has changed. How Wikipedia can win back tomorrow’s readers and how the volunteer community recognizes AI-generated Wikipedia articles was the topic of discussion at WikiCon in Potsdam in early October, the largest meeting of the German-speaking volunteer community with nearly 400 participants.
How can we reach tomorrow’s Wikipedia readers?
Current studies show that young people in particular are reading Wikipedia less and less these days. Between 2019 and today alone, visits by young people between the ages of 18 and 24 have fallen from 30 percent to eight percent. Today’s youth consume knowledge via programs such as ChatGPT or AI-generated summaries. The German-speaking communities addressed this development in a special session during the German WikiCon.
Several questions were raised: Is it due to demographic developments in Germany? Does the presentation of knowledge no longer meet the needs of young people because they have too short an attention span for longer texts? The participants in the session discussed many different aspects. In the end, it became clear that it is probably a little bit of everything.
Everyone in the room agreed on one thing: Wikipedia is one of the most important sources of freely available and verified knowledge in the world. And it is unique thanks to its structure and strong community. But new approaches are also needed to attract young people in particular as the readers of tomorrow. Developing these approaches is now one of the tasks for the future – one that Wikimedia Deutschland will tackle together with dedicated volunteers and partners from the international Movement.
How do the communities deal with AI texts in Wikipedia?
While many AI programs access Wikipedia’s knowledge, AI-generated texts are also increasingly finding their way into the encyclopedia – creating additional work for volunteers. Wikipedian Harald Krichel explained why in the session “Recognizing AI texts” at the German WikiCon.
AI-generated Wikipedia articles often contain fabricated content or references to sources that do not exist. They usually do not comply with the rules for encyclopedic writing – for example, when it comes to factual and neutral language and the balanced presentation of content. Recognizing AI texts is not an exact science, as Krichel also made clear. Nevertheless, suspicious texts can be identified with the help of so-called hard and soft factors. If such a contribution is discovered, it can be deleted or revised, depending on whether it can be adapted to Wikipedia’s rules.
These signs indicate AI articles:
- Non-existent publications or ISBN numbers
- Residual phrases from the chat, such as comments or tips from the AI
- Infoboxes that do not exist in the German-language Wikipedia
- Text fields that do not exist in infoboxes
- Sections or subheadings that should not exist in Wikipedia articles
- Web links that no longer exist in a new article
- Markup remnants such as asterisks for bold text – Wikipedia uses wikitext
- Meaningless structure, such as masses of lists, bold text at the beginning of sentences, or even emojis
- A newcomer who suddenly produces many articles
- Flowery, promotional language—AI is chatty
Reliable information in the age of disinformation
Providing reliable information – that is and remains the core of Wikipedia’s work. This requires reputable, verifiable sources. But it is precisely these that are increasingly under threat today – from disinformation, deepfakes, and the misuse of AI technologies.
This challenge was discussed at the opening panel of WikiCon. On the podium: Juliane Leopold (editor-in-chief of digital media, ARD-aktuell), Manka Heise (investigative reporter, RTL/Stern/n-tv), Tommy Krappweis (Grimme Prize winner and creator of “Bernd das Brot”) and Raimond Spekking from the supervisory board of Wikimedia Deutschland.
“The algorithmization of information has increased massively,” was one of the key conclusions of the panel. More and more often, dubious content with a wide reach is finding its way into our news feeds and “for you” pages. AI-generated clones of news sites such as tagesschau.de, fake voices, and manipulated images are being circulated—often with specific political intent. This is an attack on the foundations of democracy, which is based on verifiable facts.What does this mean for journalism – and for Wikipedia? Today, sources must be checked multiple times and information must be critically questioned. And: balance in reporting does not mean giving space to every opinion. “We don’t include every pseudoscientist and conspiracy theory in Wikipedia,” says Raimond Spekking. “Wikipedia is not about opinions,” emphasizes Juliane Leopold. The free encyclopedia thus forms a necessary counterweight to algorithmically controlled online realities. Not least because it works according to the principle of checks and balances – mutual control: “You can do that because there are so many of you!” Leopold praises the community.
The German-speaking Communities’ WikiCon is traditionally organized by a team of volunteers who are primarily responsible for the program content; Wikimedia Deutschland provides support for the planning, organization, and follow-up. Wikimedia Österreich and Wikimedia CH provide additional financial resources for the event. It usually takes place in fall (late September/early October). Presentations and recordings of several sessions from this years’ event can be found on Wikimedia Commons. The next one is scheduled to take place from September 18 to 20, 2026, in Regensburg as a hybrid event with contributions on site, online, and in combined form.
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