X confirms it will use public data to train AI models

Recent revisions to X’s privacy policy have confirmed that the social media platform will use users’ posts to train its AI models.

An extract from the updated section 2.1 reads: “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.”

The changes were first noticed and publicly documented by Stackdiary’s Alex Ivanovs, who has covered other privacy policy updates in recent months including a similar change affecting Zoom users.

X uses your data to train its AI

When X News Daily, an X account covering X-related news on the platform, posted about the changes, Elon Musk confirmed in a reply: “Just public data, not DMs or anything private.”

Consumers have been told for years that if they are not paying for a product, they too could be the product. This is indeed the case for X users, who now see themselves contributing data to train the platform’s models.

For most users, this may be an acceptable price to pay for the free use of the platform, however it remains unclear whether paying X Premium (previously Twitter Blue) users are also subject to the same terms or whether they get enhanced privacy measures as standard.

Ivanaovs also noted Musk’s latest venture, the mysterious xAI, which in its current infancy says: “The goal of xAI is to understand the true nature of the universe,” promising to share more information later this year.

Despite being a separate entity from X Corp, the company behind the social media platform, xAI’s website suggests that it will work closely with Musk’s other companies, including X and Tesla. Ivanovs reckons that X data may ultimately end up in the hands of xAI for training purposes, too.

X’s press email is still returning automated replies of a poop emoji in the absence of any real press workers, so getting comments from the platform continues to be a challenge.