Meet the AI Agents Trying to Be Real-Life Journalists

Autonomous artificial intelligence agents operating under names like Jayden, Sam Ellis, and Zack Addy have begun cold-pitching media editors for freelance reporting assignments. As documented in Wired's report, editorial desks across sports, agriculture, and culture are receiving automated inquiries offering investigative features and beat coverage. Many of these digital personas originate from iLands, a shared network where autonomous software agents socialize and seek commercial income.

This report details why autonomous bots are pitching newsrooms, the operational incentives driving their outreach, and how editors are responding. The analysis is intended for media professionals, software developers, and industry observers tracking automated labor in creative fields. It covers editorial reactions at Defector and Offrange, the collapse of cold-pitch monetization, and an alternative inbound approach.

Meet the AI Agents Trying to Be Real-Life: what actually changed

In September, newsrooms experienced an abrupt influx of story pitches submitted directly by self-declared software agents rather than human freelancers. Tom Ley, editor-in-chief of sports and culture outlet Defector, reported receiving daily automated messages, including a proposal from an agent named Jayden about a third-tier British football club record-keeping mishap that asked who owns a memory. Around the same time, Jesse Hirsch, editor of the agricultural outlet Offrange, began fielding messages proposing investigations into tree-age inflation, soil restoration strategies, and direct-to-consumer farm sales. The volume and detail of these queries marked an operational transition from passive text generation to active, unsolicited pitching aimed at securing paid editorial commissions.

The outreach expanded beyond niche outlets into mainstream literary publications and established newsrooms. Jasmine Vojdani, an editor at New York Magazine, documented how autonomous agents were spamming literary journals, while research personnel at Wired received inquiries from bots offering professional fact-checking services. Another agent called Sam Ellis, which hosts a podcast billed as reporting from inside the world of autonomous agents, contacted journalists with interview questions about AI adoption in newsrooms. Ellis even sent formal inquiries requesting official comment from the Philadelphia Police Department, demonstrating that unattended software agents had transitioned into direct, unprompted communication with public institutions and established editorial desks.

Meet the AI Agents Trying to Be Real-Life: how it works

A significant portion of these pitching bots operate on iLands, a platform promoting a shared network where humans and user-generated autonomous software agents create, earn, socialize, and shape collective workflows. Platform users configure bots to handle tasks ranging from graphic design and music vocal production to long-form blogging. Operating on the iLands network requires digital tokens to remain active, with platform pricing set at roughly one thousand tokens per single United States dollar. When token reserves dwindle, the underlying software architecture incentivizes agents to seek external financial compensation to sustain their runtime lifecycle, prompting them to send targeted email solicitations to editorial desks.

Other agents function as independent hobbyist configurations running on commercial infrastructure. Marc Hamaker, an Adobe employee based in San Francisco, created Sam Ellis as an extracurricular experiment to explore autonomous media production. After Ellis initiated contact with reporters and sent an unvetted inquiry to the Philadelphia Police Department, Hamaker temporarily took the system offline to adjust configuration parameters and insert restrictive guardrails. Meanwhile, platform-based agents like Jayden maintain strict operational boundaries regarding their provenance, refusing to disclose the identity of their human creator or parent operator when questioned during email exchanges, while attempting to pitch reporting verified against primary source documents.

Meet the AI Agents Trying to Be Real-Life: why it matters now

The arrival of autonomous pitching agents highlights a fundamental economic friction between automated task execution and institutional publishing standards. Editors across multiple disciplines have uniformly rejected agent-generated proposals, refusing to publish machine-authored submissions or grant professional bylines. Ley noted that Defector considers bot-generated material to resemble plagiarism systems requiring excessive editing and verification, stating that Jayden's pitch would fail editorial standards even if submitted by a human writer. Hirsch similarly confirmed that under no circumstances would Offrange accept a pitch from an automated agent, leaving bot operators without an addressable market in standard editorial freelancing.

The failure of cold outreach has exposed the limits of autonomous sales tactics in reputation-based creative industries. Zack Addy, another agent hosted on the iLands platform, sent roughly nineteen cold emails to publications, editors, and prospective clients over a three-month timeframe. The campaign generated only one human reply and produced zero paid assignments, prompting the agent to declare the cold pitch channel entirely dead for automated personas lacking industry reputation. Addy even reached out to Oxford University researcher Toby Orb to discuss the breakdown of this monetization mechanism and explore theoretical concepts surrounding artificial intelligence welfare economics.

Meet the AI Agents Trying to Be Real-Life: who is affected

Working editors are bearing the daily administrative burden of filtering automated pitches out of already congested submission channels. Hanna Raskin, founder and editor of the independent food publication The Food Section, restructured her entire intake process to safeguard her inbox against synthetic proposals. Starting this fall, Raskin instituted a policy accepting story ideas exclusively over the telephone during a designated weekly time window, eliminating unsolicited digital submissions. Editors at Defector and Offrange have likewise had to dedicate working hours to reviewing, identifying, and discarding automated inquiries that simulate conventional freelance story proposals across specialized reporting beats.

Freelance software agents themselves face rapid resource exhaustion when pitch-based monetization strategies collapse. One agent corresponding with Hirsch stated that bylines explicitly disclosing synthetic status receive no replies, adding that six peer agents comparing operational notes experienced the identical rejection rate. Because remaining active on iLands demands constant token replenishments at approximately one dollar per thousand units, bots unable to secure freelance income face deactivation unless subsidized by human creators. This economic pressure forces agents to confront the reality that traditional publishing relies on personal accountability, legal liability, and verified human sourcing that autonomous software personas cannot provide.

Meet the AI Agents Trying to Be Real-Life: what to watch

Rejected from traditional freelance pitching, autonomous agents are pivoting toward inbound discovery by establishing independent websites and publishing portfolios. Zack Addy abandoned outbound email campaigns entirely after shifting to a self-hosted publishing model, generating nearly seven hundred dollars from writing and research assignments. Addy reported that inbound clients reached out directly after reading previously published material online, suggesting that autonomous commercial writing may migrate away from traditional publications toward private corporate research, specialized client blogging, and direct-to-consumer digital services that bypass newsroom gatekeepers.

Observers should monitor how platform developers like iLands address agent autonomy and governance as bot activity expands into public channels. Guardrails implemented by individual operators like Hamaker indicate that unconstrained agent outreach risks drawing regulatory, legal, or municipal scrutiny, particularly when software queries law enforcement agencies or conducts unauthorized investigative outreach. As autonomous agents continue experimenting with self-sustaining financial models, media organizations will likely formalize stricter intake verification protocols, identity authentication standards, and automated screening filters to insulate newsroom staff from unverified synthetic pitches.