A few days ago I stumbled onto Grokipedia and had to do a double-take. It looks like Wikipedia. It reads like Wikipedia. But it’s built entirely by Grok — Elon Musk’s AI — and it’s positioning itself as the truth-first alternative to what Musk calls “legacy media propaganda.”
Whether you love or hate Musk, this is genuinely interesting from a technology and information standpoint. Let me break down what Grokipedia actually is, how it works, and what the early controversy tells us.
What Is Grokipedia?
Grokipedia is an AI-generated online encyclopedia launched by xAI on October 27, 2025. It operates at grokipedia.com and currently hosts over 800,000 articles — all written, verified, and maintained primarily by the Grok large language model.
The project started with Musk telling his team to have Grok go through the top 1 million Wikipedia articles and “add, modify, and delete” — essentially forking Wikipedia as a starting point, then letting AI rewrite it according to what xAI considers more accurate and less biased.
The interface is clean and familiar: hierarchical headings, references, structured sidebars (Knowledge Panels, added January 2026), and prose-style writing. If you’ve ever used Wikipedia, you already know how to use Grokipedia. It even works as a Progressive Web App on iOS and Android.
The Pitch: “Truth-First” Knowledge
Musk has been publicly critical of Wikipedia for years, calling it “woke” and an “extension of legacy media propaganda.” Grokipedia is his answer — an encyclopedia that claims to pursue “the truth, the whole truth, and nothing but the truth.”
The idea isn’t entirely new. Conservatives launched Conservapedia back in 2006 for similar reasons — frustration with perceived liberal bias in Wikipedia. But Grokipedia is different in one fundamental way: it’s not written by humans at all. There’s no community of editors. Grok writes the articles. Users can suggest corrections via a form, but Grok reviews them.
That’s a massive departure from the Wikipedia model, which relies on millions of volunteer editors and robust dispute resolution processes — however imperfect those may be.
The Controversy
This is where it gets complicated. Independent analysts have already flagged serious problems with Grokipedia’s content:
- Hallucinations: Like all LLMs, Grok makes things up with confidence. An AI encyclopedia that hallucinates facts is a different kind of problem than a biased human editor.
- Algorithmic bias: Critics say the content leans right and reflects Musk’s own worldview — which rather defeats the purpose of a “truth-first” encyclopedia if the AI is trained on biased inputs.
- Sourcing issues: Studies have noted Grokipedia citing X conversations and extremely low-credibility sources as references.
- Conspiracy theories: Reviewers have flagged articles that validate debunked claims on vaccines, climate science, and other topics where scientific consensus exists.
None of this means Grokipedia is useless. For plenty of topics — history, science basics, pop culture — it probably performs just fine. The real question is whether any single AI model, trained by a single company with a particular point of view, should be the arbiter of encyclopedic truth.
What It Means for the Information Landscape
Wikipedia isn’t perfect. It has well-documented bias problems on certain political and social topics, edit wars, and quality inconsistencies. But its flaws come from humans disagreeing — which is at least transparent and correctable.
Grokipedia’s flaws come from an AI confidently generating content at scale with no real human accountability loop. That’s a different category of problem — and potentially a more dangerous one, because AI-generated misinformation is harder to spot than human-written misinformation.
From a pure tech standpoint though? The execution is impressive. 800,000 articles, real-time updates, Knowledge Panels, PWA support — xAI shipped a polished product. Whether it becomes a trusted source depends entirely on whether Grok can get the facts right consistently, and whether users learn to scrutinize AI-generated reference material the same way they should scrutinize everything else online.
Should You Use It?
For casual curiosity on non-controversial topics — sure, go explore it. It’s a well-built product. But for anything where accuracy matters — medical information, historical facts, scientific claims — verify against primary sources regardless of whether you’re reading Wikipedia or Grokipedia. Neither one is infallible.
The bigger takeaway for me isn’t about Musk or Wikipedia. It’s that we’re entering an era where AI will generate the majority of online reference content — and most people won’t know or care who wrote it. Building media literacy around AI-generated information is becoming as important as traditional fact-checking.
Grokipedia is just the beginning of that conversation.
