X has placed Grok AI training at the center of a fresh privacy dispute. The platform, formerly Twitter, introduced a setting that lets user data connected to Grok be used for training and fine-tuning, and users were opted in by default.
The change has drawn attention because it was implemented without prior notice, according to X users who found the setting. It also comes as European regulators are already scrutinizing how major platforms use personal data to develop AI systems.
What X changed for Grok AI training
The setting allows X to use user interactions, inputs, and outputs with Grok for training and fine-tuning purposes. The collected data can also be shared with Elon Musk's AI company xAI, which trains the models powering Grok.
That means the privacy issue is not only about whether people use Grok. It is also about whether their activity with the AI assistant can become part of the data pipeline for future model development.
The setting is enabled by default for all users. The source article says it was introduced without prior notice, based on X users who discovered it.
For users who do not want their data used this way, the current practical problem is access to the control. The opt-out exists on the website, but the source describes it as hidden. The mobile app does not offer a way to opt out, though X is working on a mobile app option.
How users can opt out right now
At the moment, the available opt-out path is through the desktop version of X. The source article says users can disable the setting on the website through the Grok Settings option.
The important point is that the choice is not presented as an advance permission request. Users were automatically included, and those who object must find and change the setting themselves.
For anyone concerned about AI data use, the immediate steps are simple in principle:
- Use the web version of X rather than the mobile app.
- Open the Grok Settings option.
- Disable the setting that allows Grok-related data to be used for training and fine-tuning.
That desktop-only limitation matters because many people primarily use X through the mobile app. Until the mobile option is available, users who do not use the website may not have a direct way to change the setting from their usual device.
Why regulators are paying attention
According to the Financial Times, the Irish Data Protection Commission is investigating X's decision to automatically share user data with xAI to train its systems. The Irish authority is responsible for enforcing EU data protection rules.
The regulator had been questioning X about its plans for months and says it was "surprised" by the rollout. The Irish authority could now launch a GDPR investigation, which could lead to fines.
The central issue is consent. X users have been opted into data sharing without their explicit consent, and the source says the setting can currently only be changed in the desktop version.
The source article also notes that this kind of data collection likely conflicts with European data protection laws. It places X in a broader argument about whether platforms can rely on default settings and difficult opt-outs when training AI systems on user data.
The Meta comparison
X is not the only company facing pressure over AI training and user data in Europe. Meta recently paused plans for its "Meta AI" assistant in Europe after objections from Irish data protection authorities.
Meta's issue also involved using European user data for AI training, with automatic opt-in and a complicated opt-out process. Privacy advocacy group Noyb, which initiated the case against Meta, sees a clear violation of GDPR in using public user data for AI training without explicit consent, based on "legitimate interest".
Meta expressed frustration with the decision. The company argued that without European data, it could only offer an inferior product not optimized for local cultural nuances. Meta also said its future multimodal AI models won't be released in Europe because of "regulatory uncertainties."
That comparison helps explain why X's rollout is likely to be closely watched. The dispute is not only about one product setting. It is about the rules that will shape how AI assistants are trained, especially when platforms already hold large amounts of user activity.
What this means for X and Grok
X's actions will likely intensify the debate about using user data to train AI systems. The timing is especially notable because Grok 2 is scheduled for release in August, with Grok 3 following in December.
Musk claims Grok 3 will be the most powerful AI ever created. That ambition makes the data question more important, not less. More capable AI systems need training and fine-tuning, but regulators and users are increasingly focused on whether that work is being done with clear permission.
For X users, the practical takeaway is direct: anyone who does not want Grok interactions, inputs, and outputs used for AI training should check the desktop Grok Settings option. For X, the larger question is whether default enrollment without explicit consent can survive regulatory scrutiny in Europe.