Social media platform bans AI training on user data

Mastodon, a decentralized social network, has updated its rules to stop user data from being used to train AI models. This change comes shortly after X, owned by Elon Musk, made a similar move. The new terms, starting July 1, clearly state that scraping user data for AI training or other unauthorized purposes is not allowed. Mastodon’s email to users emphasized, “We don’t permit training AI models with user data from our servers.”

The rules also ban automated tools like bots or scrapers that collect data, except for standard search engines or browsers. These changes only apply to Mastodon’s main server, mastodon.social. Other servers in the fediverse—a network of independent servers—might still allow data scraping if their rules don’t forbid it. Platforms like Reddit and OpenAI have also added similar restrictions to protect user data.

Alongside this, Mastodon is raising its minimum user age to 16 worldwide, up from 13 in the U.S. Mastodon’s move shows growing concerns about how social media data is used for AI. While it aims to protect users, the rule’s impact depends on other servers adopting similar policies. For now, Mastodon is taking a stand to keep user data safe from AI developers.