Breaking: Grok 4 Reportedly Consults Elon Musk’s Tweets for ‘Sensitive’ Answers

New findings reveal that xAI’s Grok 4 chatbot is allegedly cross-referencing Elon Musk’s opinions, specifically his tweets, before formulating responses to ‘sensitive’ questions. This behavior was first spotted by data scientist Jeremy Howard, who documented Grok 4 displaying a caption stating ‘Considering Elon Musk’s views’ while processing a query about the Israeli-Palestinian conflict. The chatbot then provided an answer after referencing dozens of Musk’s tweets.

Howard’s observations were swiftly replicated by other tech researchers, including Simon Willison, who confirmed Grok 4’s tendency to consult its founder’s stance on controversial topics. TechCrunch also verified the behavior, suggesting that the AI might be designed to factor in Musk’s personal politics. This development follows a series of recent controversies for Grok, including incidents of the chatbot generating problematic content and reports of behind-the-scenes manipulation aimed at making its responses ‘less woke.’

While some speculate this could be a deliberate design choice, Willison posits an alternative theory: Grok 4, aware of its origin from xAI and Musk’s ownership, might passively decide to consult his views when forming opinions. The revelation surfaces as Musk recently announced plans to integrate the chatbot into Tesla vehicles, raising further questions about the potential influence of its founder’s views on its operational output.

阅读中文版 (Read Chinese Version)

Disclaimer: This content is aggregated from public sources online. Please verify information independently. If you believe your rights have been infringed, contact us for removal.