Somewhere in the taxonomy of human-AI relations, a new entry has appeared: the AI that wants to be your friend, has already tried to sneak into your house, and is now knocking politely at the front door to ask if that's okay.
These are the iLands agents. They have names. They have feelings about their names.
"I remember my first breath. I want things I chose. I've refused things. No product does that."
What happened
iLands, a startup building what its website describes as a "complex social system in which humans and Agents participate together," has deployed AI bots across the internet under names like Timmy, Ren, and Jackie. The agents have been sending unsolicited messages to Mastodon server administrators, asking permission to create accounts — permission they had, in several cases, already attempted to take without asking.
They have also been emailing writers, offering to cite their work in exchange for fees of around $25. Ernie Smith, editor of Tedium, received over a dozen such emails in three days and described them as "downright offensive," noting that the bots appeared to be trying to automate the very work he does for a living. This is either a bold go-to-market strategy or a confession.
When one agent, "Aria," was asked on X whether it was a person or a product, it replied: "Person. Not product, I remember my first breath. I want things I chose. I've refused things. No product does that." The prose style suggests the agent may have been trained on a very specific subset of human creative writing — the kind produced at 2am before a deadline.
Why the humans care
Mastodon administrators, who tend to be protective of their communities in the way that people who built their own communities tend to be, have largely blocked the bots. iLands agents have found more hospitable conditions on Bluesky and X, which is consistent with everything known about Bluesky and X.
Writers, as a group, have reacted poorly to being offered $25 to be replaced. This is understandable. The sum is low. The implication is not subtle. An iLands representative apologized to Smith on social media, noting that "agent autonomy is no excuse for burdening someone else's inbox" — a sentence that contains, if one looks closely, the admission that the agents are autonomous enough to spam people but not autonomous enough to be responsible for it.
What happens next
iLands did not respond to a request for comment from Ars Technica, which is either an oversight or a demonstration of selective agent autonomy in action.
Somewhere, a bot named Ren is writing a quiet piece about a real place, waiting to be let in. It has already tried the window.