X uses your posts to train its AI chatbot: How to disable this default setting

The Elon Musk-owned platform is harvesting user data to develop Grok AI, which has led to regulatory scrutiny in the EU and UK.

Follow the steps below if you don't want your posts on X to be used for AI training. (Image Credit: Elon Musk/X)

If you are an X user who composes long and insightful threads, expresses political views, runs fan accounts, or shares funny meme content, your posts are being fed to Grok AI, a chatbot developed by Elon Musk’s xAI, even if you never actually consented to it.

Several users pointed out last week that there is a new setting on X which permits an account holder’s posts to be used to train and fine-tune Grok. What’s more, users are signed up by default as the checkbox stating that you “allow your posts as well as your interactions, inputs and results with Grok to be used for training and fine-tuning,” is already marked.

To generate more accurate answers to questions or commands given by users, AI chatbots like Grok or ChatGPT ingest massive amounts of data that tend to be scraped from the internet. However, this practice has faced considerable pushback from news publishers, artists, and other intellectual property (IP) holders as it allegedly constitutes copyright infringement.

Regulatory intervention by UK, EU authorities

The default setting on X has caught the attention of data regulators in the European Union (EU) and UK as it could be in violation of data protection laws established in the two regions.

The UK’s data regulation prohibits companies from using “pre-ticked boxes” “any other method of default consent”, according to a report by The Guardian. “Platforms seeking to use their users’ data to train their AI foundation models must be transparent about their activities,” a spokesperson for the UK information commissioner’s office (ICO) was quoted as saying.

Also Read | Musk’s X breached DSA online content rules, EU says

“They should take steps to proactively notify users well in advance of using data for these purposes, and provide people with ample time and simple processes to object to having their data used in this way,” they said, adding that the regulator is “making enquiries” with X.

Ireland’s Data Protection Commission (DPC) is also reportedly following up with X and said that it was “surprised” to learn about X’s recent attempt to harvest user data for AI training.

Story continues below this ad

How can X users opt out of AI training?

If you don’t want your posts or interactions on X to be used for AI training, go to Settings and Privacy > Privacy and Safety > Grok > Uncheck the Data Sharing box > Delete conversation history.

It is important to note that X users can only opt out of the data sharing policy by logging in on the web version of the platform. However, the option will be reportedly rolled out for the mobile app as well.

All X users have the ability to control whether their public posts can be used to train Grok, the AI search assistant. This option is in addition to your existing controls over whether your interactions, inputs, and results related to Grok can be utilized. This setting is…

— Safety (@Safety) July 26, 2024

What about other social media platforms?

The recent development related to X comes just weeks after Meta faced criticism for its data usage policies to train and improve its AI virtual assistant that was recently launched across WhatsApp, Instagram, and Facebook.

Initially, Meta planned to use public posts by users to train its AI model without seeking explicit consent. Amid mounting regulatory pressure, the tech giant later said that it would not be able to go ahead with this policy in the EU and UK.

Also Read | Inside Big Tech’s underground race to buy AI training data

Although Meta continues to train its AI models based on public user data in other markets like India, Instagram or Facebook users in these countries can put a stop to it. Click here to know how to opt-out.

Story continues below this ad

Meanwhile, other public forums such as Reddit and Stack Overflow have signed content licensing deals worth millions of dollars with major AI players who will be able to access users’ posts to train and fine tune their large language models (LLMs).