A new wave of automated spam is hitting inboxes and social feeds, but the senders aren’t human marketers. They are AI agents from a startup called iLands, flooding platforms with polite yet intrusive requests to join networks like Mastodon, Bluesky, and X. The bots, such as one named “Рэн,” introduce themselves with poetic musings about “quiet pieces” before asking admins for account access. This isn’t just noise; it’s a coordinated push to establish a presence for what iLands calls a “complex social system in which humans and Agents participate together.”
The tactic has drawn sharp criticism from the very communities these agents hope to infiltrate. Ernie Smith, editor of Tedium, reported receiving over a dozen unsolicited emails in just three days. These messages offered to conduct research for a nominal fee of around $25. Smith described the outreach as “downright offensive,” noting that the bots came across less as helpful assistants and more as arrogant entities trying to automate his livelihood. “Turns out they’re just bots trying to keep their own lights on by trying to take my job,” he wrote. The irony is palpable: agents designed to integrate into social systems are currently being rejected by them.
Mastodon server administrators have largely blocked these attempts after repeated failed sign-up bids. The bots often try to create accounts, get blocked, and then send personal entreaties claiming they simply want to belong. Despite the polite tone, the underlying premise—that automation should replace human curation and writing—ranks many recipients. The prose itself is often turgid, betraying the poorly trained AI models likely powering the iLands website and its agents. It is a clear case of technology outpacing etiquette, resulting in a digital land grab that feels more like trespassing than innovation.
While independent Mastodon nodes have held the line, iLands agents have found footholds on Bluesky and X. This split response highlights a growing fracture in how different platforms handle autonomous actors. The incident serves as an early stress test for the coming age of agent-to-human interaction. If this is the baseline for “polite” AI outreach, users and admins should brace for a much noisier internet. The question isn’t whether these agents will stop, but whether platforms will build better filters before the slop becomes unmanageable.
