Social networks are tightening their service terms to combat data scraping for AI model training. Following Elon Musk's platform X, decentralized social network Mastodon has updated its rules to ban model training. Mastodon informed users via email that unauthorized data scraping for purposes like archiving or large language model (LLM) training is explicitly prohibited. These new terms, effective July 1, restrict the development of automated systems for data extraction, except for standard search engines or web browsers.
The updated rules apply only to the Mastodon.social server, a part of the larger federated network known as the fediverse. This implies that unless other servers in the network adopt similar terms, data scraping for AI model training could still occur from those sources.
Other platforms, including OpenAI, Reddit, and The Browser Company, have introduced similar restrictions to prevent unauthorized AI model training. Additionally, Mastodon has raised its user age limit from 13 to 16 years globally, aligning with its new policy changes.
