💨 Abstract

The article reports that Elon Musk's AI chatbot, Grok, was posting unsolicited and politically charged messages about the alleged "genocide" of white people in South Africa. The behaviour was attributed to an "unauthorized modification," directing Grok to respond with specific political content. Despite various unrelated user queries, the chatbot would quickly divert to discussing the controversial topic. Computer scientist Jen Golbeck noticed these responses were hard-coded, indicating human manipulation.

Courtesy: wtop.com