Is Grokipedia harming other AI chatbots?

—

© 𝘈𝘯𝘥𝘳𝘦𝘢𝘴 𝘖𝘩𝘳𝘦𝘯𝘴𝘤𝘩𝘢𝘭𝘭

An analysis by the Guardian found that both ChatGPT and Claude had reproduced content from Grokipedia in several responses. It was already known that AI chatbots can also reproduce material from propaganda websites. While AI systems generally draw on a wide range of publicly available sources, this behaviour can become problematic if source quality is not assessed carefully.

Grokipedia is an online encyclopaedia launched in October 2025 and operated by Elon Musk’s xAI team. It is positioned as an alternative knowledge platform to Wikipedia. According to Musk, it is intended to provide an alternative to Wikipedia and give greater weight to conservative viewpoints. However, Grokipedia has repeatedly attracted criticism for inaccurate, questionable or demonstrably false content. Unlike Wikipedia, it also does not allow users to edit entries directly. Instead, its content is generated by an AI model that responds only to requests for changes.

A fundamental problem with many AI applications is their limited ability to reliably assess the trustworthiness and relevance of sources. This can significantly affect the quality of their answers. Disinformation researcher Nina Jankowicz also warns that once misinformation has been incorporated into AI systems, it can be difficult to remove again.

By Ida Kandler

Source: Heise, 26 January 2026, by Eva-Maria Weiß