A major privacy mishap hit Elon Musk’s xAI: hundreds of thousands of Grok chatbot conversations were indexed and surfaced in Google Search.

The issue stems from Grok’s “Share” button, which generates unique public links to conversations. These links ended up being crawled and indexed by search engines. As a result, highly sensitive exchanges became discoverable - including chats about hacking crypto wallets, explicit role-play scenarios, and even instructions for making banned substances.

This comes just a month after a nearly identical incident with ChatGPT, where public links to shared conversations also appeared in Google’s index. At that time, OpenAI called it a “short experiment.”

The Grok leak raises serious questions about how AI companies handle user privacy, especially when chat histories can be so easily exposed to the open web.