Grokipedia
Last reviewed
May 31, 2026
Sources
12 citations
Review status
Source-backed
Revision
v1 · 1,648 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
May 31, 2026
Sources
12 citations
Review status
Source-backed
Revision
v1 · 1,648 words
Add missing citations, update stale details, or suggest a clearer explanation.
Grokipedia is an online encyclopedia created by xAI, the artificial intelligence company founded by Elon Musk. It launched on October 27, 2025, and its articles are generated and maintained by xAI's Grok AI model rather than written by human volunteers. Musk presented the project as a less biased rival to Wikipedia, which he has often accused of leaning to the political left. The first public release was labeled version 0.1 and shipped with roughly 885,000 articles. Reporting at launch found that many entries were copied or closely adapted from Wikipedia, some of them still carrying Wikipedia's open license notice, and the site drew criticism over accuracy, bias, and the absence of an open, transparent editing process.[1][2][3]
Musk has criticized Wikipedia for years. He has called it "Wokepedia," argued that its coverage of contested political topics reflects the views of its volunteer editors, and at one point urged his followers to stop donating to the Wikimedia Foundation, the nonprofit that hosts Wikipedia. He has also tied the idea of a public knowledge base to xAI's broader stated mission, which the company describes as an effort to "understand the universe." In that framing, an encyclopedia that an AI system can read, write, and correct fits alongside Grok and xAI's other products.[1][4]
The name combines "Grok," xAI's chatbot, with the "-pedia" suffix shared by Wikipedia and many reference works. Musk had floated the idea publicly for months before launch and described the planned product as a large improvement over Wikipedia. When version 0.1 went live he posted on X that it was "better than Wikipedia imo" and that "version 1.0 will be 10X better."[1][2]
Grokipedia went live on Monday, October 27, 2025, after Musk had teased and then delayed an earlier release date. The site opened to the public as version 0.1 with about 885,000 articles, a figure that several outlets reported and that Musk amplified on X. For comparison, the English edition of Wikipedia held more than 7 million articles at the time, so Grokipedia launched with a small fraction of Wikipedia's catalog despite the large absolute number.[2][3][5]
Each article is produced by Grok, and pages carry a note indicating that Grok generated the text. The interface is spare. Readers search or browse to an entry and read a single AI-written version of the topic, sometimes with citations to outside sources appended by the model. There is no public talk page, no visible edit history of the kind Wikipedia maintains, and no community of named editors.[1][2]
The clearest difference is who, or what, writes the content. Wikipedia is written and edited by volunteers who debate changes in the open and leave a public record of every revision. Grokipedia entries are generated by an AI model, and the public cannot edit them directly. Instead, readers can flag a problem or suggest a change through a button, and those suggestions are reviewed before any update is made. Musk and xAI described this as a system in which Grok fact-checks and improves articles over time.[1][2]
The table below summarizes the main contrasts as reported at launch.
| Feature | Grokipedia (v0.1) | Wikipedia |
|---|---|---|
| Operator | xAI, a private company | Wikimedia Foundation, a nonprofit |
| Who writes articles | Generated by the Grok AI model | Human volunteer editors |
| Public editing | No direct editing; users suggest changes for review | Anyone can edit most articles directly |
| Edit history | Not publicly visible at launch | Full public revision history |
| Article count at launch | About 885,000 | More than 7 million (English) |
| Launch date | October 27, 2025 | January 15, 2001 |
| Content license | Mixed; some pages marked CC BY-SA from Wikipedia | CC BY-SA |
These distinctions matter for how each site handles disputed facts. On Wikipedia, disagreements play out through sourcing rules and visible discussion. On Grokipedia, the model decides what to write, and the basis for a given claim is not always traceable, which is part of why reviewers questioned how readers should judge reliability.[1][6]
One of the first things reporters noticed was how much of Grokipedia read like Wikipedia. Several outlets found entries that matched Wikipedia text word for word or close to it. The Verge reported that pages such as the one for the PlayStation 5 lined up with their Wikipedia counterparts, and Ars Technica noted that Grokipedia's own article about Wikipedia was largely lifted from Wikipedia's article about itself.[1][6]
Some of those pages carried a notice stating that the content was "adapted from Wikipedia" and licensed under the Creative Commons Attribution-ShareAlike 4.0 license, the same CC BY-SA license Wikipedia uses. That license does allow reuse, including commercial reuse, but it sets conditions. Reusers must give attribution and must release their adaptations under the same license. Commentators raised questions about whether a project pitched as a corrective to Wikipedia was leaning on Wikipedia's volunteer-written text, and whether every page that drew on that text met the license's attribution and share-alike terms.[1][3][6]
Coverage of the launch was largely skeptical, and the criticism fell into a few areas. These points come from reporting by outlets including Ars Technica, NBC News, The Verge, Wired, The Guardian, Slate, and The Atlantic, and they are presented here as those outlets' findings and assessments.
On accuracy, reviewers warned that Grokipedia inherits the known weaknesses of large language models, including the tendency to state wrong information confidently. Wired framed its coverage around the idea that Grokipedia carries the same problems as Grok itself, and several writers noted that an encyclopedia with no visible sourcing trail makes those errors harder to catch and correct than on a site with open edit histories.[5][7]
On bias, the sharper complaint was that Grokipedia's framing of politically charged topics tilted toward Musk's own views or toward right-leaning positions. Reporters pointed to entries on subjects such as slavery, George Floyd, transgender topics, the January 6 Capitol attack, and COVID-19, arguing that some gave weight to contested or fringe positions or downplayed mainstream accounts. Critics also noted that articles touching on Musk himself tended to be favorable. Slate described the project as part of a wider effort by Musk to shape public information, and The Atlantic and The New Republic published similarly critical assessments. xAI and Musk have countered that Grokipedia is meant to reduce bias rather than introduce it, and that early versions will improve.[1][2][8][9]
A third line of criticism concerned transparency. Without a public edit history or an open community, observers said it was difficult to know when an article changed, who or what changed it, or why a particular claim appeared. That opacity, they argued, sits awkwardly with the project's stated goal of delivering "the truth."[1][6]
The Wikimedia Foundation responded by defending Wikipedia's volunteer model. A foundation spokesperson said that Wikipedia's knowledge is created and maintained by people, with transparent rules and open discussion, and noted that even Grokipedia drew on that human-written content. Jimmy Wales, Wikipedia's co-founder, took a measured tone in interviews, saying he was not especially worried about Grokipedia as a competitor and pointing to the reliability problems that AI-generated encyclopedias face.[2][10]
Reaction outside the encyclopedia world split along familiar lines. Musk's supporters welcomed an alternative they viewed as freer from what they see as Wikipedia's slant, while many journalists and researchers treated the launch as a case study in the limits of AI-written reference material. The recurring observation across coverage was the irony that an "anti-Wikipedia" relied so visibly on Wikipedia's own text.[3][6]
Grokipedia is one of the most prominent attempts to build a general encyclopedia mainly from AI output rather than human editing, and it arrived as generative AI systems were already reshaping how people look up information. Its launch sharpened a set of questions that extend well beyond one website. Can a large language model maintain a reference work at scale without a human review layer, and how should readers weigh entries whose sourcing is not visible. What happens to attribution and licensing when AI systems ingest openly licensed human work and republish it. And whether an encyclopedia controlled by a single company, and shaped by its owner's stated views, can earn the kind of trust that Wikipedia built over more than two decades.[2][6]
Those questions were unresolved at launch, and Grokipedia itself was explicitly a version 0.1 product that Musk promised to expand. Its early reception suggests that the speed and breadth of AI generation do not by themselves settle the harder problems of accuracy, neutrality, and accountability that any encyclopedia has to answer.[1][2]