Wikipedia is 25 years old in 2026 and has established itself as the great surviving project of the early Internet. Born in Black 2001 by Jimmy Wales and Larry Sangeralthough on the direct predecessor, Nupedia, it started appearing a year ago.
When Wikipedia started on its journey, Google was just a new browser that managed to beat Altavista, Nokia was almost synonymous with the mobile phone, and the operating system that people started accessing the Internet with was Windows XP.
Indeed, the Internet has seen many revolutions: YouTube has changed the consumption of content; electricity torrent puso en jaque a la industria cultural; games online if normalized; social leaders initiated a new form of relationship; The iPhone brought the Internet to the wallet and e-mail streamingwith Netflix at the top, redefined the audiovisual world.
However, among all these changes, Wikipedia follows ahí, like a dinosaur from the story of Augusto Monterros. Y did it without investors, without umbrellas of great companies and without publicity. How is that possible?
The answer is in its functional form. Wikipedia was born with the mission of collecting knowledge in an open and accessible library for anyone, both to consult and to add or edit information. Anyone can participate, but not everyone pays: content must be protected por fuentes fablesrated by the community volunteers you discussed low regulation claras.
In addition, some users are democratically elected moderate these processes. All content is published with open licenses (such as CC BY-SA 4.0), allowing reuse and adaptation, even for commercial purposes.
This system has allowed Wikipedia not only to survive numerous context changes over the past 25 years, but also to maintain itself as one of the most visited sites and with the highest number of users in all countries. Today, however, there is a distinct scenario before us: the disruption of artificial intelligence. ?How is Wikipedia dealing with this new digital revolution?
From human traffic to bot traffic
In recent years, the Wikimedia Fundación Wikimedia, responsible for managing Wikipedia, Wiki Commons, Wikidata and other related projects, has announced the development cases of 10% traffic due to AI generalization.
Marshal Miller, chief product officer of the Wikimedia Foundation, said that “search engines are using more than just generative artificial intelligence to provide answers directly to people who are looking for them, instead of extending to sites like ours. And younger generations are looking for information about them.” social video platform on the open web”.
No embargo, while ignoring people’s visits, the number of robots is increasing. According to Wikimedia, most large linguistic models are taught with data from Wikipedia. The result is paradoxical: the encyclopedia has to take on more of the cost of feeding the AI systems, which in turn will reduce the traffic of people dependent on its donations.
To correct this imbalance and prevent the viability of the project from being threatened, Wikimedia was launched en 2021 Wikimedia Enterprisea version of the Wikipedia API adapted for commercial use and intended especially for AI companies. Its paying customers include Microsoft, Meta, Amazon, Perplexity, Mistral AI and Google.
AI generated content
Another source of concern for those responsible for the project is the presence of AI-generated content on their Wikipedia. If this is a critical point, the value of the encyclopedia lies in the work of tens of years of thousands of human volunteers. When incorporating text produced by LLM, if future models are fed with lower quality materialswhich is an effect that various studies have shown to reduce value and quality.
This caused the nacimiento Wiki Project AI Cleanup, a volunteer initiative which is dedicated to reviewing the encyclopedia to meet and explore AI-generated content. This includes the efforts of another group of Wikipedians who have created a guide so that anyone who wants it can contribute to identifying this content. Ironically, the same guide is used to train bots so they learn to “humanize” their articles.
for a little while Elon Musk recently announced the launch of Grokipediaa great alternative to Wikipedia, built entirely from Grok’s training, on its own language model.
Considering the media impact of the announcement, however, the studies seem to indicate this Grokipedia is essentially a copy of Wikipedia content. The only saving grace is that he incorporates ideological modifications into the “sensitive” topics of the South African tycoon by expressing an opinion different from the scientific consensus.
Will Wikipedia rise to the occasion with these new difficulties? My opinion is definitely yes. Primarily because you have an integrated community of tens of thousands of people around the world compromised by your existence. Second, because it is funded through donations and without publicity, it remains one of the few independent resources on the Internet. Finally, for its greatest skill, the AI needs its own Wikipedia to continue growing.
The most likely future is the coexistence of both sources of information in a relationship that seems more like symbiosis than competence. When we’re 25 years on the Internet, Wikipedia will still be there.
***Rafael Conde Melguizo He is an investigator and professor at UDIT and a member of Wikimedia España.
Leave a Reply