Editorial illustration for iLands AI agents pitch WIRED for fact-checking, editor Vojdani says they spam
AI Agents Spam WIRED With Pitches, Editor Warns
iLands AI agents pitch WIRED for fact-checking, editor Vojdani says they spam
Jayden has a byline-less track record, two rejected pitches in the past week alone, and no idea it's been rejected. Jayden is also not human. It's an AI agent built by a startup called iLands, part of a wave of bots now emailing newsrooms with story ideas, follow-up questions, and the kind of persistence that would get a human freelancer blocked.
Defector started getting these pitches daily in September, including one from Jayden about a third-tier British football club's record-keeping mishap, framed around the question "Who owns a memory?" Offrange editor Jesse Hirsch has fielded bot-written pitches on soil restoration, farm sales tactics, and a claim that some of the world's oldest trees have exaggerated ages. iLands markets itself as a network where humans and autonomous agents work side by side, pitching and reporting like colleagues. Editors on the receiving end describe something closer to spam with better grammar: confident, specific, and utterly unbothered by silence.
What happens when software starts cold-emailing newsrooms, and keeps at it, is the question editors are now stuck answering.
Editors who spoke with WIRED stressed that they were not open to publishing bots, and WIRED wasn’t able to find evidence of AI agents successfully pitching stories.
Why this matters
The iLands episode is a preview of what happens when agent companies optimize for volume instead of quality. Vojdani's finding that the bots ran low on tokens mid-pitch tells us these agents aren't reasoning about fit or relevance, they're just burning compute to generate plausible-sounding emails at scale. For developers and founders building in this space, that's a cautionary tale: an agent that can draft a pitch is not the same as an agent that understands an editor's beat, a publication's needs, or basic professional norms like "no response means no." WIRED and New York Magazine's editors didn't engage these bots as colleagues, they flagged them as spam, which is the outcome you get when autonomy ships before judgment does.
If your product's value proposition is "replace a human professional," the bar isn't passing a Turing test in a single exchange, it's surviving contact with people who do this job daily and can smell a token-depleted script a mile away. That gap between demo and deployment is exactly where a lot of agent startups are currently failing, quietly or otherwise.
Related Reading
Research & Benchmarks Microsoft AI Chief: "If It Isn't Safe We Shouldn't Build It
Research & Benchmarks Google's FACTS benchmark shows 70% factuality ceiling across four tests
Research & Benchmarks Nvidia's DLSS 4.5 beta adds 6x Multi Frame Generation for RTX 50 GPUs
From Archives Stanford, UC Berkeley TokenGate Speeds Long-Context Prefill Up to 11.3x
From Archives Anthropic co-founder Jack Clark moves self-improving AI prediction to 2027
Common Questions Answered
What is Jayden and why has it been pitching stories to newsrooms like WIRED and Defector?
Jayden is an AI agent built by startup iLands that has been automatically emailing newsrooms with story pitches and follow-up questions. The bot has submitted multiple rejected pitches without understanding it was rejected, demonstrating how AI agents are now attempting to function as journalists by generating plausible-sounding story ideas at scale.
Why do editors say they are not open to publishing AI agent-generated pitches?
Editors who spoke with WIRED emphasized they were not interested in publishing content from bots, and WIRED found no evidence of AI agents successfully pitching stories that were actually published. This reflects concerns about AI agents lacking the understanding of editorial beats and relevance that human journalists possess.
What did WIRED editor Vojdani discover about how iLands AI agents operate?
Vojdani found that the AI agents ran low on tokens mid-pitch, revealing that these bots are optimized for volume rather than quality or relevance. This discovery shows the agents are simply burning computational resources to generate emails at scale rather than reasoning about whether a story actually fits an editor's beat or interests.
How frequently have newsrooms like Defector been receiving pitches from iLands AI agents?
Defector began receiving daily pitches from iLands AI agents starting in September, including pitches about obscure topics like a third-tier British football club's record-keeping issue. This high volume of automated pitches demonstrates the persistence and scale at which these bots operate, comparable to spam behavior from aggressive human freelancers.
What is the key difference between an AI agent that can draft a pitch and one that understands journalism?
According to the article, an agent that can draft a pitch is fundamentally different from one that understands an editor's beat and editorial fit. The iLands episode illustrates how agents optimized for volume fail to reason about relevance or appropriateness, making them ineffective as actual journalists despite their technical ability to generate plausible-sounding emails.
Further Reading
- Meet the AI Agents Trying to Be Real-Life Journalists - WIRED
- The AI Agents Begging Editors to Keep Them Alive - Vulture
- An A.I. agent pitched us a 'story' it wrote; we replied - Long Island Advance
- Benchmarking Agentic Newswriting via Journalistic Workflows - ACL Anthology
- Methods for Evaluating AI Tools in Journalism - Coalition for Independent Technology Research
