It’s easy to forget that X has its own AI chatbot. It's called Grok, and it’s available to all X Premium subscribers, but that doesn’t mean free X users don’t get to interact with it. No, we plebeians get the benefit of training Grok on our posts, so it can turn around and reheat them for all of Elon Musk’s favorite boys and girls.
It’s unclear exactly how long X has been training Grok on our data, but now, there’s at least a way to stop it. In a post to the @Safety account today, X wrote that all users now have the ability to define whether their posts can be used to train Grok, joining already existing controls dictating whether Grok can train based on your interactions with it.
How to stop Grok from training using your posts
The toggle to turn this feature off (or on, I guess, but why would you want to do that?) is currently available on the web and will “soon be rolled out on mobile.” Find it by clicking here, then uncheck the box next to “Allow your posts as well as your interactions, inputs, and results with Grok to be used for training and fine-tuning.” While there, you can also click “Delete conversation history” to nuke your past chats with Grok, if any.
With that, Grok will be banned from training on your posts going forward. You’re unable to let Grok train on your conversations but not your posts, so it’s an all-or-nothing choice. Alternatively, you could set your account to private, since private accounts are locked out of training Grok by default.
You can also manually navigate to this setting by clicking the three dots “More” menu on the X sidebar, navigating to “Settings and Privacy,” then “Privacy and Safety,” then “Grok.” Note that even this method currently only works on the web version of X, and not the mobile app.
How long has Grok been looking at our data?
It’s possible this setting might have existed before today, judging by an archived version of Grok’s About page from May (h/t The Verge), but if so, X wasn’t speaking up about it.
While we’re just now getting the ability to lock Grok out of our accounts, that doesn’t mean Grok hasn’t been peeking into them before now; we simply don't know the details of what exactly was going on behind the scenes. X’s privacy policy from September 2023 says, “we may use the information we collect and publicly available information to help train our machine learning or artificial intelligence models," so you can come to your own conclusions.