Selective divergence between Grokipedia and Wikipedia articles

Abstract

The launch of Grokipedia, an AI-generated encyclopedia developed by xAI, was presented as a response to perceived ideological and structural biases in Wikipedia, with the goal of producing more “truthful” entries using the Grok large language model. However, whether such an AI-driven alternative can systematically correct the biases and limitations of human-edited platforms remains unclear. Here we conduct a large-scale computational comparison of 17,790 matched article pairs drawn from the 20,000 most-edited English Wikipedia pages. We find that Grokipedia pages are longer, more syntactically complex, and contain fewer references per word. Similarity measures across the two platforms reveal a bimodal structure: many Grokipedia articles closely resemble their Wikipedia counterparts, while a considerable subset diverges. Political bias differences emerge primarily within the divergent subset, where Grokipedia shows a relative rightward shift in the ideological orientation of frequently cited news media sources, particularly in articles related to religion and history. These patterns indicate selective, topic-specific divergence rather than a uniform debiasing of Wikipedia content. More broadly, AI-generated encyclopedias may depart from established editorial norms by favoring narrative expansion over citation-based verification, raising questions about transparency, provenance, and the governance of knowledge in automated information systems.