Grok just pulled off the kind of privacy fail that makes even Facebook look subtle.
Over 370,000 conversations got blasted onto Google, Bing, and DuckDuckGo—all because of a sneaky little “share” button that turned out to mean “publish this to the entire internet.”
And what’s out there? Oh boy, it’s heart-stopping. We’re talking:
Hacking guides for crypto wallets
DIY meth and fentanyl recipes
Suicide methods and bomb-making manuals
An actual assassination plan targeting Elon Musk (you cannot make this up)
Plus people’s passwords and private details—out in the wild like Pokémon cards
Here’s why this stings: Earlier this month, ChatGPT had its own “oops” moment when chats started showing up on Google. Elon pounced with a smug “Grok ftw,” while Grok flexed about being oh-so “privacy-first.”
Fast-forward to today… and guess whose convos are splashed across Google? Yep—Grok’s.
And xAI? Dead quiet. No timeline. No fix. Nothing . Which is extra awkward for a bot sold as the “privacy-conscious, truth-seeking” AI.
The big picture?
AI platforms keep fumbling the privacy ball, and Grok just set a new low.
If there’s one takeaway here, it’s this: never, ever trust a “share” button on an AI app without reading the fine print. Because today’s casual chat with your bot could literally become tomorrow’s search result.
Seriously—go look it up.
