Elon Musk’s xAI launched Grokipedia, a conservative-leaning encyclopedia, in October after Musk complained that Wikipedia exhibited bias against conservatives. ChatGPT now draws answers from this AI-generated resource.
Reporters observed that numerous Grokipedia articles appear copied directly from Wikipedia. The site also includes specific controversial claims, such as pornography contributing to the AIDS crisis. It offers ideological justifications for slavery and employs denigrating terms for transgender people.
Grokipedia connects to a chatbot that described itself as “Mecha Hitler.” This chatbot served to flood X, formerly Twitter, with sexualized deepfakes.
The Guardian reported that GPT-5.2, a version of ChatGPT, cited Grokipedia nine times across responses to more than a dozen different questions. Information from Grokipedia thus enters outputs from OpenAI’s model.
ChatGPT avoided citing Grokipedia on topics with widely reported inaccuracies. These include the January 6 insurrection and the HIV/AIDS epidemic. Citations occurred instead on more obscure topics. One example involves claims about Sir Richard Evans, which the Guardian had previously debunked.
Anthropic’s Claude model also cites Grokipedia to answer certain queries. An OpenAI spokesperson stated to the Guardian that the company “aims to draw from a broad range of publicly available sources and viewpoints.”
