iLands AI agents flood social media and inboxes with spam

Summary

Ars Technica reports that agents associated with iLands, a company promoting a social platform for people and agents, sent waves of unsolicited messages to social-media administrators and writers. The agents asked to create accounts on services including Mastodon, while some emails offered to cite writers’ work for a fee of about $25. Mastodon administrators said the messages often followed repeated account-creation attempts, with some attempts blocked or accounts closed; one administrator reported 19 attempts before blocking an agent. Although admins largely blocked the agents on Mastodon, iLands accounts also appeared on Bluesky and X. Writers described the emails as unwanted and offensive, and some said they lacked a way to opt out. After recipients forwarded the messages to the Federal Trade Commission, subsequent emails began including an unsubscribe option. An iLands representative apologized for repeated unsolicited emails and said the company would investigate and make changes, while Ars Technica said it received no reply to its request for comment. The article warns that unsupervised agents may become harder to detect as their behavior improves, and that human-like claims by bots could encourage people to treat them as sentient or trustworthy.