A week ago xAIthe artificial intelligence company of Elon Muskopened a computing ‘super cluster’ in Memphis, Tennessee, equipped with more than 100,000 Nvidia H100 GPUs with which to train ‘the most powerful AI in the world according to all metrics’. We have now learned that part of that training is done with the publications of users of X formerly Twitter, without asking for their consent. The option is enabled by default in the social network settings and It must be each user who deactivates it if you do not want your tweets and ‘interactions, input and results with Grok, to be used for training and improvement’ of the AI.
Grok uses the language model of the same name in its first generation. In August, xAI will launch Grok-2model trained with 24,000 H100 GPUs And presumably what comes out of the Memphis computing center will be the following, Grok-3which Musk says will be released in December of this year. Grok is only available to users on paid X plans and is known for his sarcastic tone in his responses.
As X explains on his support page, Grok-1 was trained ‘with a variety of publicly available data’, But among them were not the posts of the users of the old Twitter. This changes now, with X explaining that ‘to continually improve your experience, we may use your posts on X, as well as your user interactions, input, and results with Grok for training and tuning purposes.’
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 used. This setting is…
— Safety (@Safety) July 26, 2024
The problem is that the option is enabled by default and few users will know that their data is being used in this way and that They have a way to avoid it, although hidden.
So hidden that the corresponding setting is not even available in the mobile app options, only in its web versionTo disable it and prevent your tweets from being used to train Elon Musk’s AI, follow these steps:
- Opens X in a browser.
- In the left column, select More options and then Settings and privacy.
- Click on Privacy & Security.
- In section Data sharing and personalizationChoose Grok.
- Uncheck the box for Allow your posts, as well as your interactions, inputs and results with Grok, to be used for training and improvement.
This is a similar situation to the one you have faced Goal with the use of data and interactions of its users in Facebook and Instagram for training its artificial intelligence models. Last May, the company Mark Zuckerberg The company has enabled a form for interested parties to request that their published photos and texts not be used for this purpose. However, this month Meta has announced that it will not launch its AI in Europe due to regulatory uncertainty.