xAI Merger Raises Questions About Grok Training Despite EU Ban

Key Takeaways

• Elon Musk’s AI company xAI recently acquired X for $33 billion.
• The merger could make it easier fro xAI to train Grok on X user data.
• However, the Irish Data Protection Authority has already restricted X’s AI training.

When Elon Musk announced that xAI had acquired X for $33 billion, he said the move would “combine the data, models, compute, distribution and talent” of the previously separate companies.

Specifically, the merger could make it easier for xAI to access X’s vast wealth of user data. But in the EU, regulatory hurdles may still prevent the AI startup from using X data to train its chatbot, Grok.

GDPR and Social Media AI Training

After xAI started training Grok on X posts in 2024, the Irish Data Protection Commission quickly intervened to stop the practice, arguing that such data processing was in violation of the EU’s General Data Protection Regulation (GDPR).

The Musk-owned company isn’t the only social media platform caught up in debates over AI training and GDPR.

Meta has also been forced to suspend similar programs in Europe after privacy advocates filed a string of GDPR complaints.

X Privacy Concerns

By bringing Grok development in-house, X will be able to harvest training data without requiring user consent to transfer data to a third party.

A spokesperson for the data protection NGO, noyb, which initiated the GDPR complaints against Meta, told CCN that with the two companies united, “all of X’s user data now belongs to xAI, turning X’s ‘might share’ into an ‘already has.’”

“Musk’s merger has just turned X into a huge pool of AI training data,” noyb said,

“It is almost certain that X’s user data will be used in its entirety for AI training, making data protection on X a thing of the past” they added.

How privacy watchdogs respond to the merger remains to be seen.

So far, European regulators have been sympathetic to the view expressed by privacy advocates—that any AI training with personal data requires an explicit opt-in.

But the issue remains contested, with Big Tech firms such as Meta lobbying for a more expansive interpretation of GDPR rules.

AI Developers Navigate European Regulation

With stricter privacy requirements compared to the U.S., navigating European regulations can be a challenge for AI developers.

In 2024, Meta excluded the EU from its multimodal Llama rollout, blaming the “unpredictable nature of the European regulatory environment.” Despite those concerns, eight months later, the company lifted its EU embargo.

For AI developers, there is an easy way around GDPR rules. They simply need to gain consent from data subjects.

But platforms have repeatedly discovered that when data processing is opt-in rather than opt-out, participation rates plummet, with many more users declining to hand over personal data.