ChatGPT’s latest AI model, GPT-5.2, has been found sourcing its information from another AI-generated website, xAI’s Grokipedia, raising concerns about the model’s accuracy.
According to The Guardian, GPT-5.2 cited Grokipedia as a source nine times when the outlet asked it questions about some lesser-known topics, including the Iranian government’s ties to telecom company MTN-Irancell and British historian Richard Evans. (Anthropic’s Claude was also found citing Grokipedia for some queries.)
Grokipedia was built by Elon Musk’s xAI startup to take on Wikipedia, a platform he thought was biased. Whereas Wikipedia’s pages are put together by volunteer human editors, most of Grokipedia’s 6,092,140 articles are generated by Grok, though readers can submit edit suggestions. When Grokipedia debuted, we found that many entries were copied from Wikipedia or paraphrased from other sources, too.
At issue is a problem that researchers have been warning about for quite some time. If AI chatbots get their information from unreliable sources, it could lead to a lot more disinformation.
Grok has been caught spreading misinformation on multiple occasions. Additionally, a report from last year said that a Russia-based disinformation network was trying to manipulate models like ChatGPT and Grok by publishing millions of articles that push their own narrative in the hopes that the AI would scrape them and use the information in their answers.
When asked about GPT-5.2’s use of Grokipedia as a source, an OpenAI spokesperson told The Guardian that the model “aims to draw from a broad range of publicly available sources and viewpoints,” adding that they already have a system that filters out low-credibility information.
Recommended by Our Editors
We asked GPT-5.2 some random questions about Evans and the Iranian government, but found no references to Grokipedia in its responses. This suggests OpenAI may have limited the use of Grokipedia as a source, or that it appears only for some specific prompts.
Disclosure: Ziff Davis, PCMag’s parent company, filed a lawsuit against OpenAI in April 2025, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.
Get Our Best Stories!
Your Daily Dose of Our Top Tech News
By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use and Privacy
Policy.
Thanks for signing up!
Your subscription has been confirmed. Keep an eye on your inbox!
About Our Expert
Experience
Jibin is a tech news writer based out of Ahmedabad, India. Previously, he served as the editor of iGeeksBlog and is a self-proclaimed tech enthusiast who loves breaking down complex information for a broader audience.
Read Full Bio