Grokipedia selectively drawing on more-right leaning news sources, new study

Posted on: 19 May 2026

A large-scale analysis of Grokipedia, the world’s first AI-written encyclopedia, has found that while many Grokipedia articles closely resemble their Wikipedia counterparts, a substantial subset diverged markedly in style, sourcing, and political leaning.
 
Conducted by researchers at Trinity College Dublin and Technological University Dublin, the study compared nearly 18,000 of the most-edited English-language Wikipedia pages with articles on the same topic on the new Grokipedia platform.
 
The study is the largest academic analysis of Grokipedia since it was launched by Elon Musk last October with a promise that the AI-written encyclopedia systematically “fixes” left-leaning biases alleged to exist in the widely used online encyclopedia Wikipedia.
 
Wikipedia’s content is written and maintained by volunteer editors, while Grokipedia is an AI-generated encyclopedia using the xAI’s Grok large language model.
 
What did the study find?
 
Using computational text analysis and machine learning methods, the team analysed articles on the same topic across Wikipedia and Grokipedia. Selection of topics was based on Wikipedia’s most-edited English-language pages. The team compared differences in writing style, structure, and the political orientation of external sources referenced in the paired articles.
 
The researchers found a profound split – while many Grokipedia articles closely mirror Wikipedia, a substantial proportion (66%) of the 18,000 analysed are more extensively rewritten – they are longer, more complex, and rely on fewer references.
 
As a whole, articles on Grokipedia show similar political leaning to those on Wikipedia, drawing on left-leaning news sources. However, when it comes to the politically and culturally sensitive topics of religion, history, literature and art, Grokipedia shows a consistent shift toward referencing more right-leaning news sources compared to Wikipedia.
 
The study analysed Wikipedia’s most-edited English-language pages, a selection that likely overrepresents high-profile and contentious topics. That said the study, according to the authors, provides useful evidence of emerging differences between AI-generated and human-edited encyclopedic knowledge systems.
 
Details of the research, conducted at the joint Centre for Sociology of Humans and Machines (SOHAM) in Trinity and TU Dublin, have been published in the peer-reviewed journal Proceedings of the National Academy of Sciences (PNAS). 
 
What is the impact of this research?
 
Lead author of the study, Saeedeh Mohammadi, PhD candidate at SOHAM and Research Ireland’s Centre for Research Training in Foundations of Data Science said:  “Online encyclopedias are central to public knowledge. They are also being used to train future generations of large language models. Our findings raise important questions about how public knowledge is produce, reproduced, verified, and governed.
 
“Unlike Wikipedia, where biases are visible and contested through human editing, AI-generated systems operate largely opaquely. This means shifts in perspective or sourcing may occur without clear accountability or editorial oversight. Simply put AI generation does not remove bias – it changes how and where bias enters the system, often making it less visible.”
 
Professor Taha Yasseri Director of SOHAM and Principal Investigator of the study said: “Rather than systematically ‘correcting’ Wikipedia’s alleged biases, as claimed when first launched, our findings suggest that AI-generated encyclopedias such as Grokipedia selectively reshape existing knowledge. This creates a patchwork system in which some content is copied, while other content is reinterpreted in ways that are less transparent and harder to scrutinise.”
 
“There is a dire need for transparency, oversight, and regulation in this space. Our information landscape is changing rapidly. We have already seen how the lack of editorial responsibility on social media platforms has enabled the generation and circulation of misinformation and disinformation, often with catastrophic consequences for elections, public health, and social stability. Now, we are witnessing the large-scale, black-box regeneration of information by large language models that remain largely closed to public scrutiny.”
 
The publication, ‘Selective divergence between Grokipedia and Wikipedia articles’ in Proceedings of the National Academy of Sciences (PNAS) can be read in full on the journal website. The publication has emanated from research supported in part by grants from Research Ireland.

• Tags:
• Culture
• Research
• Society