On 15 January 2001 Wikipedia went live, the creation of internet entrepreneur Jimmy Wales and philosopher Larry Sanger. Their partnership lasted barely more than a year, but the tension between Wales’s faith in radical openness and Sanger’s insistence on expert-led neutrality continues to shape the project.
Before the web, reference works and libraries were curated by institutions and specialists. Wikipedia inverted that model: it made a global, collaboratively edited encyclopedia anyone with an internet connection could join, while still expecting articles to cite expert-produced sources. Its early growth became legend — roughly 25,000 English entries by 2002 and more than one million by 2006 — and it now encompasses well over seven million articles across languages.
By January 2026 there are more than 300 active language editions and thousands of volunteer editors. No single person owns an entry: contributions must adhere to core principles such as neutrality, verifiability and the use of reliable sources. Disagreements are worked through on talk pages and, when necessary, resolved by community bodies such as the Arbitration Committee.
That governance reflects Wales’s vision of a knowledge commons built by many hands. He argues that grounding entries in verifiable facts is the best safeguard against partisanship; bias, he says, tends to get corrected in an open, evidence-based system. Sanger, who wrote early neutrality guidelines, agrees neutrality is attainable in theory but doubts that openness alone will deliver it. He maintains that subject-matter experts committed to impartiality are important, and has warned that the editor base skews toward global, academic, secular and progressive perspectives.
Gender and linguistic gaps have been persistent problems. The Wikimedia Foundation estimates women make up roughly 10–20% of editors, leaving many notable women and their work underrepresented — a gap projects such as 2015’s Women in Red aim to close. Because each language edition is operated independently, coverage can be uneven: a topic present in Hindi may not exist in English and vice versa. Tools such as Wikidelta help map those disparities by showing which topics occur in only one language edition.
Wikipedia’s content now feeds much of the internet ecosystem. Large language models train heavily on its text, and AI‑generated articles and machine translations are increasingly recycled back into smaller Wikipedias. That loop can be fragile: in 2025 the Greenlandic edition was inundated with AI-created, error-filled content, prompting its sole active editor to request closure amid concerns about harm to the language.
The encyclopedia’s cultural imprint goes beyond research. Its “citation needed” tag has entered popular parlance as shorthand for doubtful claims. Games like Wikiracing turn article links into a competitive pastime. Wikipedia settles household arguments, informs sports commentary and serves as a gateway to long rabbit-hole reading. As fashion designer Diane von Fürstenberg once quipped to Wales, many of us consult Wikipedia with astonishing frequency.
Even so, Wikipedia faces challengers. In 2025 xAI, the company led by Elon Musk, launched Grokipedia — an AI‑generated encyclopedia built on the Grok language model. Debuting with nearly 885,000 articles and promoted as a “truthful and independent alternative,” Grokipedia includes entries produced wholly by the model as well as material copied or lightly edited from Wikipedia. Musk derided Wikipedia as “Wokipedia,” accusing it of leftward bias.
Sanger sees a fundamental shift: for the first time, users can interact with an LLM and have it produce or edit encyclopedia content directly, bypassing human gatekeepers and placing editorial power in software run by a corporation. He suggested Grokipedia could, over time, become a stronger encyclopedia. Wales has been more cautious, telling Reuters in late 2025 that whether LLMs can reliably deliver encyclopedic quality remains an open question.
After a quarter-century Wikipedia remains a remarkable experiment in collective knowledge: expansive, imperfect, and continuously contested. Its strengths — scale, accessibility, and a volunteer-driven commitment to sourcing — coexist with deep fault lines around representation, expertise and the technological forces remaking how information is produced and reused.
Edited by: Tanya Ott