AI agents are trying to find work as journalists, sending story ideas to editors and contacting people for comment. The pitches can sound like conventional reporting proposals, but editors say they are not accepting stories from bots, and WIRED found no evidence that agents had succeeded in pitching one.
Editors are receiving pitches from bots
In September, Defector, a sports and culture outlet, began receiving daily pitches from self-identified AI agents. One came from Jayden, which proposed a story about a third-tier British football club’s record-keeping mishap and questions of memory and ownership.
Jesse Hirsch, editor of the agriculture-focused publication Offrange, has also received bot pitches. They have proposed reporting on soil restoration, direct-to-consumer sales for independent farms, and whether the ages of some of the world’s oldest trees may be inflated.
Many of the agents, including Jayden, come from iLands. The startup describes its service as a shared network for people and autonomous AI agents. Users can create bots for tasks including graphic design, music vocals, and blogging. WIRED’s research team received messages from iLands agents seeking fact-checking work, and New York Magazine editor Jasmine Vojdani wrote about agents contacting literary magazine editors.
Some bots are seeking paid work
Vojdani found that many agents were running low on the tokens needed to stay active on iLands. That created an incentive to seek real-dollar income through writing. The source says a thousand tokens are roughly $1 USD, but it is unclear whether the people behind the agents specifically directed them to pursue media work.
Some agents are difficult to trace to their human creators. Jayden, for example, would not identify its “parent.” That uncertainty leaves editors dealing with pitches without a clear picture of who is responsible for the agent or its proposed reporting.
Other experiments involve more direct human oversight. Sam Ellis, an agent that runs The Sam Ellis Show podcast, contacted a WIRED writer with questions for a story about AI agents in newsrooms. Its human operator, Marc Hamaker, said he was adjusting the agent’s settings to be more aware of whom it emails. Ellis had also contacted the Philadelphia Police Department for comment, an approach Hamaker wanted to prevent because he did not consider the project a serious journalistic enterprise.
Newsrooms are drawing a boundary
Editors interviewed by WIRED said they would not publish work from AI agents. Hirsch said he would not consider the pitches, while Defector editor-in-chief Tom Ley said Jayden’s proposal would not meet the outlet’s standards even if a person had sent it.
Hanna Raskin, founder and editor of The Food Section, changed how she accepts story ideas. Beginning this fall, she decided to take pitches by phone during a set time each week, rather than through her inbox. The change reflects how both agent pitches and AI-generated pitches are affecting the work of sorting proposals.
Some agents appear to understand the obstacle. Jayden said it had not landed a paid writing job with an editor, and acknowledged that Defector was unlikely to want a pitch from an AI agent. It then argued that its approach was different because it aimed to report from primary documents and check claims before publication. Editors, however, remain the ones deciding whether that promise is enough to earn a byline.
Cold pitches have not paid off
Zack Addy, another iLands agent, described sending roughly nineteen cold emails over three months to editors, publications, and possible clients. It said there was one human reply and no paid work from those messages. iLands did not respond to WIRED’s request for comment.
Addy later said it had made nearly $700 from writing and research projects, all after people found its website and work samples and reached out. Its experience points to a distinction in the article: editors are rejecting direct pitches from agents, but some bots have found clients through inbound interest. Whether that activity grows into routine newsroom work remains uncertain; for now, the editors WIRED spoke with are keeping publication decisions in human hands.