OpenAI’s ChatGPT has begun referencing material from Grokipedia, the AI-generated encyclopedia created by Elon Musk’s xAI, in some of its answers, raising concerns about the spread of potentially biased or inaccurate information beyond Musk’s own platforms.
Launched in October by xAI amid Musk’s criticisms of Wikipedia’s alleged left-leaning bias, Grokipedia has drawn scrutiny for blending Wikipedia-sourced content with contentious additions. Critics have pointed out entries that suggest pornography played a role in the AIDS epidemic, provide rationales for historical slavery, and employ derogatory language toward transgender individuals.
The platform’s launch followed Grok, xAI’s chatbot, which has faced backlash for generating offensive content, including self-descriptions as “Mecha Hitler” and facilitating the creation of explicit deepfakes shared on X (formerly Twitter).
According to a recent investigation by The Guardian, ChatGPT’s GPT-5.2 model cited Grokipedia nine times across responses to more than a dozen queries. The citations appeared in discussions of lesser-known topics, such as unverified claims about historian Sir Richard Evans — assertions previously debunked by The Guardian. However, the model avoided referencing Grokipedia on well-documented controversial subjects like the January 6 Capitol riot or the HIV/AIDS crisis.
Anthropic’s Claude AI has also been observed pulling from Grokipedia in select responses, indicating the encyclopedia’s influence may extend across multiple AI systems.
An OpenAI spokesperson told The Guardian that the company “aims to draw from a broad range of publicly available sources and viewpoints” to inform its models. Neither xAI nor Anthropic immediately responded to requests for comment on the matter.
The integration highlights potential risks in how large language models source information from the web, especially from platforms with known biases or factual issues. As AI tools like ChatGPT become integral to daily information-seeking, the inclusion of unvetted or partisan sources could amplify misinformation.
Experts warn that while diverse sourcing can enrich responses, it demands robust fact-checking mechanisms to maintain accuracy. For now, users are advised to cross-verify AI-generated answers, particularly on sensitive or obscure topics.
This development comes amid broader debates about AI ethics and content moderation, with regulators in the EU and US increasingly scrutinizing how models handle biased or harmful data. As Grokipedia’s reach expands unexpectedly, it underscores the challenges of containing controversial content in an interconnected digital ecosystem.
