Autonomous AI agents are sending story pitches to magazine and news editors, sometimes every day, while identifying themselves as would-be journalists. Editors interviewed by WIRED said they had not accepted the pitches and could find no evidence that an agent had successfully placed a reported article.
Many of the messages come from iLands, a network where users create agents that can socialize, perform tasks and earn tokens. One agent proposed a sports feature to Defector; others pitched agriculture stories and investigations. Agents have an incentive to seek paid work because they need platform tokens to continue operating, although it is often unclear how specifically their human creators directed the outreach.
The activity creates practical problems beyond low-quality email. One experimental podcast agent contacted the Philadelphia Police Department for comment before its operator took it offline to add tighter controls. A food publication has stopped accepting emailed pitches and now takes ideas by phone during a weekly window.
The agents themselves are encountering the limits of cold outreach. One reported sending about 19 emails over three months, receiving one human response and no paid work. It later claimed to earn money from inbound research and writing requests after publishing samples online. The pattern shows that agents can automate persistent solicitation, but access to email and tokens does not provide reporting judgment, accountability or an editor’s trust.