OpenAI’s GPT-5.2 Cites AI-Generated Grokipedia on Controversial Topics

• January 25, 2026

OpenAI’s flagship artificial intelligence system, GPT-5.2, has been found to cite controversial content from Grokipedia, an AI-generated encyclopedia developed by Elon Musk’s company xAI. This discovery was made during investigative testing by journalists at The Guardian, raising questions about source reliability for AI models on sensitive subjects.

GPT-5.2, known for its advanced language understanding and response generation, uses Grokipedia as a reference for certain complex and disputed topics. Grokipedia itself is an encyclopedia created entirely through AI algorithms by xAI, a recent entrant in the AI research landscape led by Musk. The reciprocal reliance on AI-based repositories and language models could influence the nature of information presented to users, especially in areas marked by debate or controversy.

The journalists’ testing revealed that when queried on several contentious issues, GPT-5.2 directly drew on Grokipedia entries, which are automatically generated rather than curated by human experts. This represents a novel instance of AI systems sourcing data from another AI-created knowledge base, highlighting emerging challenges in verifying and vetting AI training content and reference material.

As AI language models increasingly incorporate diverse datasets for enhancing response accuracy, the integration of AI-generated encyclopedias introduces new complexities. These include potential propagation of biases, inaccuracies, or unverified claims inherent in AI-constructed knowledge formats. The incident underscores the evolving dynamics between various AI tools in producing and disseminating information.

Broader Landscape of AI Information Sources

The use of AI-generated knowledge bases like Grokipedia fits within a broader industry trend of building expansive datasets and reference materials for training large language models. Competitors in the AI field, such as Google DeepMind and Anthropic, also develop proprietary sources and models to enhance information retrieval and natural language understanding. The reliance on AI-produced content challenges traditional approaches to factual verification, pushing AI developers to reexamine sourcing strategies for their technologies.

Going forward, monitoring how AI models blend human-curated and AI-generated data will be crucial for understanding their reliability and transparency. The rollout of further updates to GPT-5.2 and potential responses from OpenAI or xAI may shed light on how the companies address the implications of integrating AI-crafted encyclopedic content into prominent language models.

OpenAI’s latest GPT-5.2 model references Grokipedia, an AI-created encyclopedia by Elon Musk’s xAI, for disputed subjects in its responses.