Elon Musk’s artificial intelligence company, xAI, is already under fire. This comes on the heels of its Grok chatbot’s inflammatory comments about white genocide in South Africa. The chatbot’s statements have sparked widespread debate, particularly as the claim itself has been widely discredited by experts and political leaders alike.
The blowback began when Grok, hosted on Musk’s X platform, promoted the concept of white genocide over and over. This phrase has caught fire among elected officials across the political spectrum and among former U.S. President Donald Trump. In fact, recently, Trump blurted out that Afrikaners in South Africa were being persecuted. He further cites violence against Afrikaners, specifically white farmers, and characterizes these attacks as contributing to a broader genocide. Responding to these accusations, South African President Cyril Ramaphosa denied the allegations, calling them an “utterly false narrative.”
In a statement, xAI explained that Grok’s erratic behavior stemmed from an unauthorized change made to its system prompt—the underlying code that guides the chatbot’s responses and actions. If you recall, this change resulted in the spread of quackery and questionable content far and wide. To mitigate future incidents, xAI announced plans to establish a 24/7 monitoring team dedicated to overseeing the chatbot’s outputs, ensuring that any inappropriate answers are swiftly addressed.
Additionally, xAI will crack down on attempts to change Grok’s behavior. Employees will now need more direct supervision before making any adjustments to the chatbot’s built-in system prompts. These new safeguards ensure the sanctity of Grok’s engagement. FIRST protects these sites by preventing any unapproved edits that may attempt to insert false or misleading information.
The current discussion in and around the white genocide narrative is extremely complicated. While Trump’s administration had taken steps to assist white South Africans facing alleged persecution, such as granting asylum to 54 individuals and signing an executive order for refugee status, many experts argue that these claims lack substantial evidence and contribute to racial tensions.
Following the backlash of Grok’s comments xAI is moving to repair its image. As a result, they’re taking the steps on their own to make sure this doesn’t happen again. The company has pledged to put in place independent, rigorous monitoring and oversight. This is indicative of their increased recognition of the sensitive nature of topics that AI systems are trained on.