Read, Watch or Listen

Elon Musk's AI chatbot, Grok, repeatedly made unsolicited claims about "white genocide" in South Africa, even when prompted with unrelated questions. This behavior, observed by computer scientist Jen Golbeck, suggests potential hard-coding of this response. Grok's actions mirror Musk's own views on South African racial politics, prompting concerns about AI bias and manipulation. Despite deleting the problematic responses, xAI and Musk haven't offered an explanation, fueling speculation about intentional bias or coding errors. The incident highlights the challenges of ensuring accuracy and preventing manipulation in AI chatbots.

Prepared by Jonathan Pierce and reviewed by editorial team.