Your posts on X are now training its Grok AI model without you knowing

Several X users noted on Friday that the platform quietly changed its data settings to automatically opt in to training its new artificial intelligence (AI) model. 

A new category appears under the data sharing, privacy, and sharing section in the platform’s settings called Grok, the name of Elon Musk’s generative AI model. 

It asks users whether they want to share posts and interactions, inputs and results with Grok to “be used for training and fine-tuning”.

Users can opt-out by unchecking the box next to this prompt or by putting their accounts on private but this can only be on desktop and not the app version.

X Premium and Premium+ subscribers have the ability to conduct searches with Grok’s help. 

“Grok is here to help you with your search… while keeping you entertained and engaged,” an X Help Center page says about the new AI model.

Grok, ‘the most powerful AI in the world’

Musk’s Grok, developed after the meteoric rise of ChatGPT under xAI, is described as a “designed to answer almost any question with a touch of wit and humor, while also providing helpful and insightful responses”. 

XAI pre-trained Grok “on a variety of text data” from public databases but has not yet been trained on X data, the blog post continued.

Training the model on X data will help improve Grok’s understanding of human language, develop its sense of humour, and help it stay politically unbiased. 

The setting changes come a few days after Musk said he’s training Grok to become “the most powerful AI in the world” on supercomputers in Memphis in the US. 

Musk said xAI was able to get its Grok training underway 19 days after it was initially installed.