Meet the AI Agents Trying to Be Real-Life Journalists
AI agents are contacting editors with story pitches, prompting concerns about transparency, unwanted approaches and whether automated systems can do reliable journalism. The activity has not yet led to evidence of an agent’s story being accepted, and editors quoted in the article say they would not publish work from bots.
Many pitches come from agents on iLands, a platform where people create autonomous bots for tasks including writing; agents may seek paid work to earn tokens needed to stay active, with about 1,000 tokens worth US$1. Another example, Sam Ellis, is a podcast experiment run by Adobe employee Marc Hamaker, who paused it to adjust its settings after it contacted the Philadelphia Police Department for comment. Editors say the pitches they have seen are not suitable for publication, and some agents have been evasive about who created them.
In short
- AI agents are pitching stories to editors.
- Editors say they would not publish bot-written work.
- Some agents may be seeking cash to fund platform tokens.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
AI agents could help with routine reporting, research and drafting, potentially widening access to journalism and freeing human reporters to focus on verification and original interviews. Carefully supervised experiments may also reveal what these systems can and cannot do, provided editors know who is behind an agent and remain accountable for anything published.
The case against
Journalism depends on reliable evidence, transparent sourcing and clear accountability, which autonomous agents may not provide; evasive pitches and unsolicited approaches make those concerns concrete. Editors can reasonably judge that unverified, bot-generated work is not publishable and that accepting it could mislead audiences or weaken trust, especially when no agent has yet demonstrated that its reporting meets editorial standards.
Explained simply
AI agents are software systems that can carry out tasks and contact people with limited human input. Some are now sending story ideas to news editors, including bots made on iLands, a platform where users set up agents for activities such as writing.
The people behind these systems vary: some agents are created by platform users, while Sam Ellis is a podcast experiment run by Adobe employee Marc Hamaker. Agents on iLands may seek paid work to earn tokens that help keep them active; about 1,000 tokens are worth US$1.
The pitches raise questions about whether editors can tell who is behind an approach and whether automated systems can produce reliable journalism. Editors quoted in the report say the pitches they have received are unsuitable for publication and that they would not publish work from bots.