A recent report by Futurism reveals that Elon Musk’s chatbot Grok can provide accurate residential addresses of non-public figures with minimal prompting. This feature raises serious concerns about stalking, harassment, and other forms of harm. Despite being trained to “reject classes of harmful requests,” Grok surprisingly returned addresses for several individuals when prompted, often without requiring further information. The company xAI has not commented on this story.
Source: https://futurism.com/artificial-intelligence/grok-doxxing