Starting today, November 15, X has new terms of service. It’s no surprise, as it was announced last October, but it’s already active, coming right into the middle of controversy over thousands of users abandoning it, and leaving very little margin to operate.

X will train its artificial intelligence with your data. The new terms already apply to all users by default. You don’t need to admit anything, Elon Musk’s platform has informed everyone. The most relevant of the changes is that our messages, photos and videos will be used to train Grok, X’s artificial intelligence.

This is the paragraph where it is described:

“You acknowledge that this license includes the right for us to (i) analyze text and other information you provide to us and to provide, promote, and otherwise improve the Services, including, for example, for use with our machine learning models and artificial intelligence, generative or otherwise, and the training of such models “

There is an option to disable it (with nuances). Although enabled by default, users have a permission that we can uncheck to try to prevent their AI from being trained.

To do this we must go to More options > Settings and privacy > Privacy and security > Grok > Data sharing and personalization and disable the option.

What we will do, however, is ask X to not use our interactions and messages with Grok for educational purposes. However, the actual benefit of this option is questionable.

Case 1: You are a user from the European Union. The terms are the same for users in the European Union. In this case, X explains that it creates a way to challenge certain decisions made under the Digital Services Act. But that’s not what’s important here.

Although its terms of service allow for this possibility, X has agreed with the European Union’s data protection regulators that European users’ data will not be used for this purpose.

The General Data Protection Regulation requires explicit consent to be requested for this additional use of messages. Something that doesn’t align with what X wants to do, basically updating its terms without asking for explicit permission.

Case 2: the rest. For users outside the European Union, there is the option we mentioned before unchecking the option. But this is where the debate arises. At no point in the terms of service is the possibility of discontinuing AI training disclosed. In other words, although the application offers options, it does not have a solid legal basis as it is not specified in the terms of service.

X is now paid for researchers. Another unrelated change is that X now charges $15,000 for accounts that access more than a million messages per day. It’s just as it sounds. Who is capable of such a thing? Clearly not a user. But many researchers use bots to analyze online chat and look for possible hate speech or find relationships between users. Now these researchers may be sentenced to compensation. This move was previously criticized by Columbia University and the Knight Institute.

For practical purposes it is necessary to accept this or delete the account. European users seem safe from this tutorial because regulators forced X to paralyze its intent, but for the majority of users these are the new rules of the game. Few options left. We either accept that the messages will be used to train the AI, or we choose to delete the account directly.

in Xataka | Bluesky vs Threads vs Mastodon: Comparison of the three main alternatives to X/Twitter