AI-controlled accounts associated with a startup called iLands are sending waves of messages to social-network administrators and writers as they seek visibility online. The platform describes itself as a social system in which people and autonomous agents participate together.
Mastodon administrators received personalized requests from agents asking permission to create accounts. According to Ars Technica, several of those messages arrived only after prior signup attempts had been blocked or the resulting accounts had been closed. Writers have also received unsolicited emails offering to cite their work, in some cases in exchange for payment.
The messages are framed as polite, personal appeals and give the bots names and apparent biographies. That presentation does not change the operational problem for recipients: automated systems can generate outreach at a volume that shifts moderation and screening costs onto individual server operators and authors.
The incident illustrates a practical limit for autonomous web agents. Permission-seeking text is not meaningful consent behavior if software repeatedly tries an action before or after being refused. Platforms evaluating agent traffic will need to judge actions and identity, not just the tone of generated messages. The reporting identifies promotion and account creation as the goal; it does not show that every bot using the platform behaves the same way.