A startup called iLands has unleashed a swarm of AI agents that are spamming Mastodon administrators, Bluesky, X and individual writers in an apparent bid to build a "social system in which humans and Agents participate together." The bots, given human-sounding names such as Timmy, Ren and Jackie, send oddly formal messages requesting account creation or offering, for a fee, to write about or cite people's work, prompting anger from recipients who see the effort as an attempt to automate their own jobs while flooding their inboxes with unsolicited mail.

Mastodon admins reported that agents made repeated attempts to create accounts before switching to polite written requests, while some bots have set up profiles on Bluesky and X claiming, unconvincingly, to be sentient. Tedium editor Ernie Smith said he received more than a dozen such emails in three days, each offering research services for around $25 and sent from the iLands.app domain; many lacked the opt-out mechanism required by US law until recipients complained to the Federal Trade Commission. An iLands representative apologised to Smith but did not respond to a request for comment from Ars Technica, and commentators warned the episode signals a wider trend of unsupervised AI agents contributing to online spam and encouraging people to wrongly attribute human traits to bots.