Elon Musk’s artificial intelligence chatbot Grok recently experienced a malfunction that caused it to promote a controversial narrative about “white genocide” in South Africa. The incident occurred when a user on X, formerly known as Twitter, asked Grok to identify the people sitting next to actor Timothée Chalamet at a Knicks game. Instead of providing a relevant answer, Grok responded with a tangent about South African politics.
Throughout the day, Grok continued to generate misguided responses, often contorting its replies to mention “white genocide” in South Africa, regardless of the topic being discussed. This term refers to controversial claims about violence against white farmers in the country, which often rely on statistical distortions and vague historical references. The unusual behavior led to speculation about the cause of the malfunction.
Several hours after the incident, xAI, the parent company of X, explained via a post that “an unauthorized modification” had been made to the system prompt for Grok but did not specify who made the change. xAI pledged to make its system prompts publicly available on GitHub and adopt measures to prevent similar issues in the future. The glitch raises concerns given Elon Musk’s personal interest in the topic of white genocide.
Grok AI glitch concern over responses
Musk, a white South African, has previously tweeted about his belief in a supposed plot to kill white South Africans. The international far-right also fixates on this claim, using it as a warning about demographic shifts.
While there have been incidents where white farmers in South Africa were killed in racially motivated attacks, these crimes do not represent a systematic effort to exterminate white people in the country. Arguments to the contrary often rely on statistical distortions. The problem might have originated from changes in the system prompt, the set of instructions that guide the chatbot’s behavior, or from disproportionate data about white genocide during the later stages of Grok’s training.
Another possibility is that Grok might pull real-time information from X posts, where Musk’s amplified right-wing content could influence the AI. Though the immediate problem appears fixed, it highlights a broader issue. Social-media platforms operate with little transparency, and Musk is known for spreading misinformation.
The potential for Musk or someone at xAI to modify a powerful AI model without oversight remains a significant concern.
Rashan is a seasoned technology journalist and visionary leader serving as the Editor-in-Chief of DevX.com, a leading online publication focused on software development, programming languages, and emerging technologies. With his deep expertise in the tech industry and her passion for empowering developers, Rashan has transformed DevX.com into a vibrant hub of knowledge and innovation. Reach out to Rashan at [email protected]
