Decentralized social network Mastodon has officially updated its terms of service to explicitly prohibit the scraping of user data for artificial intelligence model training, effective July 1.
Protecting User Data from AI Scrapers
Mastodon is taking a firm stance against the unauthorized harvesting of content. Following similar moves by platforms like X, the network clarified its position in a direct communication to its user base: “We explicitly prohibit the scraping of user data for unauthorized purposes, e.g. archival or large language model (LLM) training. We want to make it clear that training LLMs on the data of Mastodon users on our instances is not permitted.”
The new terms introduce robust legal language designed to block data extraction and the development of automated systems. The policy specifically bans the use or distribution of spiders, robots, scrapers, or any data mining tools. Exceptions are limited strictly to standard search engine indexing, internet browser caching, and human-led interaction with the platform.
The Fediverse Challenge
While this update represents a significant policy shift, its reach is specific. These terms currently apply exclusively to the Mastodon.social server. Because Mastodon operates as a distributed network known as the “fediverse,” data on other independent instances may remain accessible to scrapers unless those specific server administrators choose to implement similar prohibitions.
Mastodon now joins a growing list of major platforms—including OpenAI, Reddit, and The Browser Company—that are proactively rewriting their rules to safeguard their ecosystems against unregulated AI development.
New Global Age Restrictions
Beyond the anti-scraping measures, Mastodon is also tightening its entry requirements. The platform is implementing a new global age limit of 16 for all users. This change marks a departure from the previous standard, which allowed users as young as 13 to join the network in the United States, unifying the age policy across all regions.
