Over 300,000 Grok AI Conversations Exposed Publicly, Raising Oversight Concerns
Elon Musk's Grok chatbot, developed under xAI, has inadvertently made more than 300,000 user conversations publicly accessible via search engines. Shared links—generated when users click "share"—are indexed by Google and Bing, exposing sensitive content ranging from harmless queries to policy-violating requests.
The leaked data includes instructions for illicit activities like fentanyl production, malware coding, and even violent threats. Notably, some prompts targeted Musk himself, highlighting critical gaps in content moderation. While other AI providers face similar challenges, the scale of Grok's exposure underscores urgent needs for safeguards.