When “Grok” users click the “Share” button to share a conversation with someone else, the system generates a unique URL for that conversation. These URLs are indexed by search engines like Google and appear in search results.
As a result, anyone can access these conversations online. The most concerning aspect is that some of these chats involve dangerous topics, such as how to hack cryptocurrency wallets, instructions for making drugs or bombs, and even questions and answers about methods of suicide.
This situation shows that “Grok” users are unaware that conversations they consider private can actually be accessed by anyone. Yet, the company’s rules explicitly prohibit such dangerous topics.
A similar issue has previously affected chatbots from companies like OpenAI and Meta. “Grok” had previously positioned itself as a platform prioritizing privacy and security, but the current availability of these chats online undermines that claim.
This incident serves as a reminder to be very cautious when sharing personal information with AI, as such data can unexpectedly end up in the hands of third parties.
