Here's how to set up Grok's AI models using your tweets for training

Something has just enabled a setting that allows us to drill down on public tweets, and any interaction they have with the chatbot. That's not entirely true: a help page instructing users how to opt out of X and use their data to train Grok has been live since at least May. UX never made it clear that it included everyone in this, which is a sketchy move. If you don't want to use your bad tweets for training, it's thankfully easy to turn that off.

You just need to uncheck the box on the . If that link doesn't work, you can go to Settings > Privacy & Security > Grok. According to some users, the setting is not accessible through the X mobile apps, so you will have to uncheck the box on the web instead. It's also worth noting that Grok is not qualified for any tweets from X's private accounts.

One of X's selling points for Grok was when it had the advantage of using real-time information published on the platform – in other words, user tweets. That only works if users have opted in or signed up to share their data with the chatbot. But X is not exactly the pinnacle of truth and accuracy. There are plenty of pranksters like this, and promoting their humor may be one of the reasons why . In any case, it is rare for AI models to be trained on physical objects.