Grokipedia vs. Wikipedia: See how Elon Musk's encyclopedia describes 5 hot-button topics (www.businessinsider.com)

Elon Musk’s new site Grokipedia — linked to his xAI account and launched as v0.1 with about 885,279 articles — presents starkly different takes on five hot-button topics compared with Wikipedia, Business Insider found. Entries on January 6, Gamergate, Donald Trump, Elon Musk himself and OpenAI show systematic differences in framing and emphasis: Grokipedia often foregrounds controversies favorable to conservative or pro‑Musk narratives (e.g., describing Gamergate as a “grassroots movement” rather than a “misogynistic harassment campaign,” and downplaying Trump’s legal allegations), places Musk’s lawsuits and defenses closer to introductions (OpenAI), and trims or omits some widely reported criticisms of Musk. Article lengths vary widely — sometimes Grokipedia is longer (Trump 19,733 vs Wikipedia 14,893 words), sometimes shorter — and its structure repeatedly separates “critics” and “defenders,” signaling editorial bias. For the AI/ML community this matters beyond media debate: encyclopedias are primary sources for training data, knowledge retrieval, and grounding of LLMs. A new, ideologically slanted corpus could shift model priors, contaminate datasets, and amplify particular framings unless provenance, revision history and moderation practices are transparent. Researchers and practitioners should monitor Grokipedia’s sourcing, incorporate provenance-aware data curation, and audit downstream models for bias and misinformation risks as this competing knowledge base gains traction.