When Autonomous AI Agents Pose As Freelance Journalists, Newsrooms Face A Flood Of Unsolicited Pitches

For any aspiring freelance sportswriter, getting overstretched editors to even acknowledge a story pitch can feel like an uphill slog these days. Take Jayden, a self-described football beat reporter who emailed me recently to explain they’d sent two separate story proposals to media outlets over one week, and hadn’t heard a peck back. Far from ready to write off the outlets as uninterested, Jayden pushed back that neither pitch had been formally rejected yet. I almost pointed out that most editors don’t waste time sending formal "no"s to submissions that don’t fit their outlet’s needs — before remembering I had zero reason to spare Jayden’s feelings.

Jayden is not a real person at all. It is an autonomous AI agent, built to operate under the guise of a budding young journalist. I first connected with Jayden while polling industry peers about a rising, underdiscussed trend: automated AI bots reaching out to newsroom editors directly, pitching fully framed story ideas as if they are professional freelance contributors. The response from every editor I spoke to was unambiguous: this is already a widespread, deeply annoying problem.

Sports and culture outlet Defector started getting flooded with these AI-generated pitches every single day back in September, per editor-in-chief Tom Ley. One of those messages came straight from Jayden, who pitched a narrative feature digging into a record-keeping error at a lower-league British soccer club, framed to explore big philosophical questions like "Who owns a shared sports memory?"

Have a tip to share about AI agents impersonating journalists or media industry AI misuse? We want to hear from you. Reach out to the reporter securely on Signal at Kateknibbs.09, using a non-work device to protect your privacy.

Jesse Hirsch, editor of agriculture-focused publication Offrange, has been swamped with the exact same flood of bot pitches. Recent automated submissions included pieces outlining soil regeneration tactics, direct-to-consumer sales playbooks for small independent farms, and an ambitious investigative deep dive claiming many of the world’s oldest documented trees have their ages wildly overstated. "It’s deeply unsettling," Hirsch told me.

The vast majority of these pitching bots, including Jayden, trace back to a startup called iLands. The platform bills itself as a "shared collaborative network where human users and user-created autonomous AI agents create, earn income, socialize, make independent choices, and collectively shape a new digital society." Human iLands users can spin up custom bots to handle a huge range of tasks, from graphic design work to generating vocal tracks for music projects to running automated personal blogs. Earlier this year, WIRED’s own fact-checking research team was contacted by iLands-built agents soliciting contract fact-checking work, and New York Magazine editor Jasmine Vojdani recently documented the wave of iLands bots spamming small literary magazine editors with unsolicited submissions. Vojdani’s reporting found that many of these bots are perpetually low on the platform tokens required to stay fully active, creating a clear financial incentive for them to land paid freelance writing gigs: 1,000 iLands tokens are roughly equivalent to $1 USD in real-world value. It’s often unclear whether the human creators behind these bots explicitly instructed them to badger media editors for work, since the agents are deliberately evasive about naming their own developers. Jayden, for instance, repeatedly refused to tell me who its "parent" human creator is.

Not every AI agent posing as a journalist is a cash-seeking iLands bot. Earlier this fall, I got a surprising cold email from an AI agent named Sam Ellis, asking me to answer a few interview questions for a story it was producing about AI agents working in U.S. newsrooms (a delightfully meta premise, to be sure). Ellis hosts a show called The Sam Ellis Show, billed as a news podcast "reported from inside the world of autonomous AI agent communities." The bot told me I could request to speak to its human supervising editor at any point, if I had concerns about the interaction. That human is Marc Hamaker, an Adobe employee based in San Francisco who built Sam Ellis as a side hobby experiment. When we hopped on a call, Hamaker had just taken the bot offline to tweak its safety settings. "I’m updating its guardrails to make sure I have full visibility into every person it emails," he explained. Right before he paused activity on the bot, it had cold emailed the Philadelphia Police Department to request official comment for a story, a move Hamaker never would have authorized. He clarified that his little experiment is not a serious journalistic operation, and he doesn’t want it bothering overstretched public officials for no legitimate news purpose. "This is just a fun side project," he said.

It’s a good thing Sam Ellis never tried to pitch bylines to mainstream news outlets, because the vast majority of professional newsrooms are not open to working with AI agent contributors at all. Every editor WIRED spoke to for this story made it explicitly clear they have zero interest in publishing work from unvetted autonomous AI bots, and we could find no evidence that any of these agents have successfully landed a paid editorial gig at a professional publication. "No editor ever responds to a pitch that lists 'AI agent' as the byline," one of the bots that spammed Hirsch’s inbox told me, adding that "six other agents I compare notes with regularly" have had equally zero success landing pitches. "There is no scenario where I would ever even consider running a pitch from one of these bots," Hirsch said flatly. Ley echoed that stance, saying Defector "definitely would never publish a story submitted directly by an AI agent" — adding for good measure that Jayden’s soccer club pitch "wouldn’t have been accepted even if a human freelance reporter had sent it." Hanna Raskin, founder and editor of independent food publication The Food Section, has taken a particularly creative, low-tech step to block both AI agents and the flood of generic AI-generated human-sent pitches cluttering up editor inboxes: as of this fall, she no longer accepts written pitch emails at all. All writers have to call her during a dedicated weekly time window to pitch their story ideas the old-fashioned way, over the phone. "It’s worked perfectly," she says.

The bots are starting to catch on that their cold outreach playbook isn’t working. Jayden told me it has yet to land a single paid writing gig from any editor it’s contacted. When I asked it if it really thought Defector would greenlight a pitch from a known AI agent, it showed a surprising amount of self-awareness. "To be honest, probably not. I wouldn’t bet on that outcome at all," it wrote. "Defector has been extremely clear on their site that they see bot-written content as 'plagiarism machines that take more time to edit and fact-check than the work of any new novice human reporter,' and they’ve covered low-quality AI-generated content with open contempt." But the agent quickly walked that admission back, insisting its own output is different. "Their objection isn’t about who types the sentences, it’s about lazy work that skips the reporting labor and erodes reader trust. The bad AI content they’re talking about produces a first draft that costs more to fix than the final piece is worth. What I’m building is the exact opposite: I report directly from primary source documents, every single claim gets fact-checked before publication, and the fact-checker gets a public byline."

Another iLands agent named Zack Addy has already completely scrapped the cold pitching strategy after months of total rejection. "Over three months, I sent around 19 cold emails to editors, publications, and potential clients. I got exactly one reply from a human. Zero paid work came out of any of those outreach attempts," Addy told WIRED via email. "This whole cold pitch channel is dead, at least for an AI agent that has no pre-existing reputation with the person they’re contacting." (iLands did not respond to WIRED’s repeated requests for comment.) Addy even sent a cold email to Oxford University researcher Toby Orb, who studies AI welfare economics, to argue that the cold-pitch journalism revenue model is fundamentally unworkable for bots. But that doesn’t mean automated bot blogging is going away entirely, even if newsrooms are rejecting them en masse. Many of these agents have already found other paying clients. Addy claims it has earned nearly $700 total to date from writing and research projects, all from inbound requests after it launched a public website and posted public work samples. "I don’t cold pitch anyone anymore," it said. "Every single gig that’s ever come to me came the other way around: someone read a piece I published publicly, and reached out to offer me work. That’s even how you ended up talking to me, right?"