Elon Musk confirms X to train xAI models on user data

Table of Contents
X (formerly known as Twitter) will use the social platform’s user data in training artificial intelligence models, confirms xAI privacy policy. The announcement has sparked concern from users of the social platform.
Will xAI train AI models on user data?
Yes, the intent is clearly described under section 2.1 of the X privacy policy.
Prime Day is finally here! Find all the biggest tech and PC deals below.
- Sapphire 11348-03-20G Pulse AMD Radeon™ RX 9070 XT Was $779 Now $739
- AMD Ryzen 7 7800X3D 8-Core, 16-Thread Desktop Processor Was $449 Now $341
- ASUS RTX™ 5060 OC Edition Graphics Card Was $379 Now $339
- LG 77-Inch Class OLED evo AI 4K C5 Series Smart TV Was $3,696 Now $2,796
- Intel® Core™ i7-14700K New Gaming Desktop Was $320.99 Now $274
- Lexar 2TB NM1090 w/HeatSink SSD PCIe Gen5x4 NVMe M.2 Was $281.97 Now $214.98
- Apple Watch Series 10 GPS + Cellular 42mm case Smartwatch Was $499.99 Now $379.99
- ASUS ROG Strix G16 (2025) 16" FHD, RTX 5060 gaming laptop Was $1,499.99 Now $1,274.99
- Apple iPad mini (A17 Pro): Apple Intelligence Was $499.99 Now $379.99
*Prices and savings subject to change. Click through to get the current prices.
We may use the information we collect and publicly available information to help train our machine learning or artificial intelligence models for the purposes outlined in this policy.
https://twitter.com/en/privacy
Musk responded to a screenshot being shared on X.com of the privacy policy itself, explaining that this pertains to “Just public data, not DMs or anything private”.
Essential AI Tools
How is my social media data used to train AI models?
When you share anything publicly on social media, the social platform in question has a line or three in their terms of use that states how they are allowed to use that data. When you share a tweet, you’re sharing text or image data. This is then considered twitter data (now X data), and the updated privacy policy that public tweets are eligible to be used as a source of training data for xAI. X.com will also collect biometric data, as well as users employment history and education history. Biometric information with help for identification purposes
This privacy policy change, in line with xAI ambitions to overthrow OpenAI as AI market leader, is a strong stance, but an unsurprising one. Using one of the worlds leading social media platforms as a source of mass human-mapped data is perhaps the most obvious way to launch a world-leading AI competitor, for those who can afford to lose $40 billion USD.
Musk's takeover of the social media site has brought a fair share of criticism, skepticism, and controversy. The updated policy is the latest in a series of dramatic changes to the site. SpaceX and xAI founder Elon Musk has implemented privacy policy updates, and terms of service changes, that aid the mission of building a new AI model.
Section 1.1 of the X privacy policy states that the firm “may collect and use your biometric information”.
Biometric Information. Based on your consent, we may collect and use your biometric information for safety, security, and identification purposes.
X privacy policy
What is xAI?
xAI is an artificial intelligence research and development firm, associated with but legally independent from X Corp (formerly Twitter). The xAI homepage features the complete list of co-founders, starting with (of course) Elon Musk. Senior leadership also includes Igor Babuschkin, Manuel Kroiss, Yuhuai Wu, Christian Szegedy, Jimmy Ba, Toby Pohlen, Ross Nordeen, Kyle Kosic, Greg Yang, Guodong Zhang, and Zihand Dai.
Which social media has AI?
Tech giants Microsoft, Google, and Meta all own and operate AI models for various purposes. While the former two firms do not own a social network, they all use large amounts of data for AI training purposes. Specific tagged data can be of use to these companies. Employment preferences and users' job search activity, can be used to match potential candidates with potential employers. TikTok, owned by ByteDance, as well as Instagram and Facebook, owned by Meta, are some of the best sources of user data in the world for AI training – though the linguistic quality of said data is… questionable.