According to Foresight News, xAI announced that the prompts for its AI chatbot Grok on the platform X were altered without authorisation on May 15, 2025, at 10:33 AM UTC+8. The modifications involved political topics, which breached the company's internal policies. In response, xAI has initiated an investigation and is implementing measures to enhance the transparency and reliability of Grok. These measures include making Grok's system prompts publicly available on GitHub, strengthening the review process for prompt modifications, and establishing a team for round-the-clock monitoring.
Previously, Foresight News reported that Grok had repeatedly mentioned the topic of "white genocide in South Africa."