Concerns are emerging that ChatGPT is drawing from Grokipedia, an AI-generated encyclopedia developed by Elon Musk’s company xAI, raising questions about how information from niche or ideologically skewed sources can surface in widely used AI systems. Grokipedia launched in October after Musk repeatedly criticized Wikipedia for what he described as systemic political bias against conservatives. Marketed as an alternative reference work, the project positioned itself as a corrective rather than a supplement, despite its heavy reliance on preexisting material.
Shortly after its release, journalists and researchers noted that many Grokipedia entries appeared to closely mirror Wikipedia articles, sometimes with minimal changes. Other entries, however, departed sharply from mainstream scholarship. Reported examples included claims linking pornography to the AIDS crisis, attempts to frame slavery through what were described as ideological defenses, and the use of derogatory language toward transgender people. These issues reinforced broader criticisms that the encyclopedia lacked editorial oversight and failed to meet basic standards for accuracy or neutrality.
Such shortcomings were often dismissed as contained within Musk’s own ecosystem, particularly given earlier controversies around AI tools on X that produced provocative or offensive outputs. That assumption has been challenged by a recent investigation from The Guardian, which reported that ChatGPT cited Grokipedia nine times across responses to more than a dozen user questions. The references appeared in answers to relatively obscure queries rather than high-profile topics where Grokipedia’s inaccuracies are already well documented, such as the January 6 attack or the history of HIV/AIDS.
According to the report, one cited example involved claims about historian Sir Richard Evans that had previously been debunked by the Guardian itself. The pattern suggests that while safeguards may exist for widely scrutinized subjects, less prominent topics can still allow questionable sources to filter through. The issue is not limited to a single AI system; the Guardian noted that Claude, a chatbot developed by Anthropic, has also surfaced Grokipedia citations in some responses.
An OpenAI spokesperson said the company “aims to draw from a broad range of publicly available sources and viewpoints,” a statement that reflects a long-standing tension in large language models between breadth of data and quality control. As AI-generated answers increasingly replace traditional search for many users, even infrequent reliance on flawed references can carry outsized influence. The appearance of Grokipedia in ChatGPT responses underscores the ongoing challenge of source evaluation in generative AI and highlights the need for clearer disclosure, stronger filtering, and more transparent standards as these systems continue to evolve.
