Grokipedia, the AI-written encyclopedia backed by Elon Musk, has not accepted or rejected a single user-suggested edit in over three months, according to a report from Lawfare. That silence raises a bigger question: can an AI-only knowledge base stay accurate without ongoing human or machine correction?
What actually happened
Grokipedia launched as v0.1 in October 2025 with 885,000 AI-generated articles. xAI released v0.2 in November 2025, growing the site to more than 6,000,000 articles, according to grokipedia.com/live. That page's "Recent changes" section now reads "No live edits available right now." The Verge reported that Lawfare found no entry changed since 24 April 2026. Lawfare reviewed 34,519 pages with at least one suggested edit, covering 225,496 recommended changes. "In our subsequent analysis... we found no accepted or rejected corrections dated within the past three months," Lawfare said. xAI did not immediately respond to a request for comment. Musk's last public mention of Grokipedia on X was in February 2026.
How we got here
Musk framed Grokipedia as a "massive improvement" over Wikipedia when xAI launched the project, promising an AI-written alternative free of what he called ideological bias. The site scaled quickly, jumping from 885,000 articles at launch to over 6,000,000 within a year, all generated and, in theory, corrected by AI systems rather than volunteer editors. Wikipedia itself relies on thousands of active human editors who process edits daily. Grokipedia's edit and correction pipeline, by contrast, appears to have gone quiet just as the platform reached its largest scale, according to Lawfare's review of tens of thousands of flagged pages.
Why this matters for you
For users, an AI encyclopedia that never updates loses the one advantage it claimed over Wikipedia, the ability to correct itself fast. Errors flagged months ago, across more than 225,000 suggested edits, remain unresolved. For builders exploring AI-generated reference tools, including those feeding data into AR and smart-glasses assistants, the lesson is blunt: automation without a maintenance layer stalls. Any product that pulls facts from Grokipedia risks serving stale or wrong information. For xAI and Musk, the silence undercuts public claims made at launch and raises questions about whether the project has enough staffing or automated review capacity to keep pace with its own article count.
The bigger question
Can an encyclopedia built entirely by AI stay trustworthy without continuous human or automated correction, or does scale without maintenance just produce a larger pile of unverified claims? Grokipedia grew from 885,000 to over 6,000,000 articles in about a year, but its correction queue appears frozen. If a six-million-article AI reference cannot keep its own edit backlog current, what does that mean for AI systems, including AR and wearable assistants, that plan to cite it as a source of ground truth?
What to watch
No date has been set for xAI to resume Grokipedia edits or respond to the Lawfare report. Watch for any statement from xAI or Musk addressing the frozen correction queue, and for whether the "Recent changes" page on grokipedia.com/live starts showing activity again. Bonuz will track how AI-generated knowledge bases handle accuracy as they get pulled into AR and smart-glasses assistants.
