X Is Quietly Training Grok With Your Posts, Here’s How You Can Opt Out

• X has switched on a setting that allows it to train Grok on user posts and interactions with the chatbot without publicly announcing it.
• Simultaneously, the platform has enabled users to opt out of the feature.

X, the popular social media platform, has turned on a setting that allows it to train an AI assistant, Grok, on millions of user posts and interactions. Elon Musk’s company, xAI, is the developer behind Grok.

The move came to light when an X user, Kimmy Bestie, with the handle @EasyBakedOven, said the social media platform had just activated a default setting for everyone. However, Musk did not publicly announce this change when he mentioned that the AI chatbot had started supercluster training.

A notice on X’s settings page says, “To continuously improve your experience, we may utilize your X posts as well as your user interactions, inputs, and results with Grok for training and fine-tuning purposes. This also means that your interactions, inputs, and results may also be shared with our service provider xAI for these purposes.”

How Grok Is Currently Trained

Most generative AI (GenAI) models are primarily trained through machine learning (ML), which allows machines to learn from data and past experiences to identify patterns. Specifically for GenAI models, ML uses text and data mining (TDM), which involves feeding large amounts of data from both open-source and copyrighted material.

Many companies developing GenAI models rely on social media platforms to train their models, as it requires massive amounts of data; much of the data on these platforms is public and free. Using social media to train AI models is also cost-effective. Furthermore, the goal of GenAI is to act and sound like humans. The more these AI models can learn from actual people posting on social media postings and their interactions, the closer they come to achieving the goal.

While companies like Meta use data from their own platforms (e.g., Facebook and Instagram) to train their models, some companies rely on third-party platforms. This has led to many companies like OpenAI, Slack, and Clearview AI coming under fire for potential copyright infringement and privacy concerns over misusing personal user data.

Like most large language models (LLMs), Grok-1 was trained on a variety of data publicly available on the internet and data sets curated and reviewed by human reviewers. However, it was not pre-trained on X data, including public X posts.

At the same time, the post on X’s help center says that Grok can decide whether or not to search public X posts and perform a real-time search on the internet. The chatbot’s access to public X posts allows it to respond to user queries with the latest information and insights on various topics.

This may imply that Grok isn’t being trained on posts and interactions from private X accounts.

See more: Elon Musk Set To Integrate xAI Startup With X Platform

EU Regulators Concerned

A European privacy watchdog, the Irish Data Protection Commission (DPC), took note of the development and said it was surprised by the platform’s move. The DPC leads on oversight of X’s compliance with GDPR, which allows for penalties for confirmed breaches of up to €20 million or 4% of global annual turnover.

Graham Doyle, deputy commissioner of DPC, said, “The DPC has been engaging with X on this matter for a number of months, with our latest interaction occurring as recently as yesterday. Therefore, we are surprised by today’s developments.” Doyle further said that the watchdog followed up and was waiting for a response from X.

How to Opt Out

Interestingly, a help page instructing users how to opt out was already available even before Kimmy Bestie wrote the post on X. X also seems to have made it easy for users to opt out of this feature. While users can’t turn it off using the X mobile app, they can do it on the website.

The following are the steps to turn off X’s data-sharing settings:

• Open the Settings pageOpens a new window on X on your laptop or desktop
• Select the “Privacy and safety” button
• Select “Grok”
• Uncheck the box

Users can also delete their conversation history with Grok at once. According to X’s help center, deleted conversations will be removed from the systems within 30 days unless they are kept for legal or security reasons.

Here is how you can delete your conversation history:

• Open the Settings pageOpens a new window on X on your laptop or desktop
• Select the “Privacy and safety” button
• Select “Data sharing and personalization”.
• Select Grok
• You will find “Delete conversation history”.
• Conform to “Delete your interactions, inputs, and results”

Visit the X Help Center hereOpens a new window to learn more about Grok and keeping your account safe.

MORE TECHNOLOGY NEWS

• Microsoft and OpenAI Incorporate Artificial Intelligence Features in Search Engines
• Mimecast Acquires Code42
• Researchers Identify Significant Vulnerabilities and Malicious Activities Within GitHub
• Threat Actors Exploit Telegram Zero-Day Vulnerability Hiding Malware in Video Files