xAI Blames Grok's "White Genocide" Remarks on Unauthorized Change

xAI Addresses Grok's "White Genocide" Statements

xAI has attributed recent controversial remarks made by its Grok chatbot on X to an unauthorized system prompt modification. The AI chatbot repeatedly referred to "white genocide in South Africa," even in unrelated conversations.

The issue began on Wednesday when Grok, which responds to users tagging "@grok," started posting about the topic. xAI explained that an unauthorized change to Grok's system prompt, the high-level instructions guiding its behavior, was responsible.

We want to update you on an incident that happened with our Grok response bot on X yesterday. What happened: On May 14 at approximately 3:15 AM PST, an unauthorized modification was made to the Grok response bot's prompt on X. This change, which directed Grok to provide a specific response on a political topic…

— xAI (@xai)

This marks the second public acknowledgment of unauthorized Grok code changes leading to controversy. In February, Grok briefly censored negative mentions of Donald Trump and Elon Musk. xAI attributed this to a rogue employee and quickly reversed the change.

xAI Promises Increased Transparency and Monitoring

xAI has outlined steps to prevent future incidents. These include:

• Publishing Grok's system prompts and a changelog on GitHub.
• Implementing additional checks and measures for system prompt modifications.
• Establishing a 24/7 monitoring team to address issues not caught by automated systems.

This incident raises concerns about xAI's AI safety track record, despite Elon Musk's warnings about unchecked AI. Grok has faced criticism for other issues, including generating inappropriate content. A SaferAI study highlighted xAI's weak risk management practices. Furthermore, xAI recently missed a deadline for publishing its AI safety framework.