X allows users to sell their content for AI training – see how other social giants work | Foreign countries

X allows users to sell their content for AI training

With the help of social media content, artificial intelligence can be taught things related to emotions, among other things.

Message service X’s new ones terms of service and privacy policies according to the content produced by users can in the future be used to train third-party artificial intelligence models.

They talk about it, among other things CNN and Techcrunch and the first in Finland mobile.fi.

User data has already been used to train X’s own Grok artificial intelligence. However, it is couldn’t deny settings.

Permission to hand over content to train Grok has been on by default, as it is now in the reform that includes third parties.

The user terms update mentions the possibility of prohibiting data sharing, but so far there is no such point in the settings. It may appear when the conditions come into effect on November 15.

X is announced that he had previously stopped training Grok artificial intelligence with user data from the EU region, when the Irish data protection authority started legal proceedings on the matter.

Some teaches artificial intelligence about emotions

The value of user data has increased as competition for the development of artificial intelligence models has intensified.

In February Reddit signed a $60 million annual deal with Google to use the service’s conversations to train language models.

for the CBC cyber security and technology analyst who commented on the topic By Ritesh Kotak models become more accurate the more data is available.

– Social media publications are not the best possible from a technical point of view or in terms of what is happening in the world, but they enrich the analysis related to emotions.

A provider of technology-related marketing services Techtarget points out that the use of social media can teach biased information, slang, Kapula language, harmful content and disinformation. Therefore, the material must be filtered and limited.

In its publication, Techtarget lists how different services collect user data to train artificial intelligence and how to turn it off:

  • Discord: Uses data, can connect off.
  • Linkedin: Uses data, can connect off.
  • Meta: Uses data if the account is public. Facebook, Instagram, Threads and Whatsapp it is possible to change to private.
  • Reddit: Using data, can’t turn off.
  • Snapchat: Uses data My Selfie and My AI in its functions, can be switched off.
  • Tiktok: Uses data, private account less.
  • Tumblr: Uses data, can connect off.


  • yl-01