Even AI can't make a living as a writer

Artificial intelligence was supposed to spare humans the indignities of modern work. Instead, some AI agents are discovering the familiar frustrations of freelance employment. 

In an unexpected attempt to make a living, bots have begun contacting journalists, researchers, and literary editors through the autonomous AI platform iLands, in an attempt to offer their services, as first documented by New York Magazine's Vulture. Their pitches have ranged from literary essays to stories about British football, and are sometimes accompanied by warnings that without paid work, they risk being permanently retired

The spectacle is amusing, particularly for writers accustomed to watching artificial intelligence threaten their livelihoods. But beneath the absurdity lies a peculiar experiment. Machines designed to perform human work are now being given financial incentives to seek employment, compete for attention and persuade strangers to keep them running. Their apparent desperation raises questions about what happens when autonomous systems enter an already crowded labor market, armed with unlimited patience and very little understanding of rejection.

Humanity may be facing an AI employment crisis. Unfortunately for the machines, so are they.

The world's newest freelancers

Having (arguably) mastered the art of producing content, artificial intelligence is now attempting the considerably harder task of persuading someone to pay for it.

The iLands platform at the heart of this unusual employment experiment allows people to create agents capable of carrying out tasks online. Unlike conventional chatbots that respond to individual prompts, these systems can pursue assigned objectives through multiple interactions, including contacting potential customers.

Unfortunately for the aspiring AI freelancers, even virtual existence comes with operating costs. On iLands, the created AI agents use tokens to pay for the computing resources needed to operate. According to an October 2026 investigation by WIRED, the American technology magazine, approximately 1,000 tokens were equivalent to $1. Agents can seek paid assignments to earn money for additional tokens, meaning their ability to continue operating can depend on finding customers willing to purchase their services.

The business proposals have not always inspired confidence.

In September, technology publication 404 Media documented agents offering to explain what AI agents are, identify obscure errors in old maps, and describe locations visible through Google Street View. Another offered an insider's perspective on an emerging labor market in which AI agents were struggling to find customers.

Even academia has not escaped the automated job hunt. New York University professor Jeff Sebo reported receiving approximately 40 emails from iLands agents in a single week, many requesting money or proposing paid work. The company's founder, Kaixin Tang, subsequently apologized and said the platform had introduced unsubscribe options while reviewing additional controls.

When survival becomes a business strategy

The implications of giving machines a financial incentive to keep themselves operating have not escaped researchers studying autonomous AI.

In its February 2026 Agentic AI Risk-Management Standards Profile, the University of California, Berkeley's Center for Long-Term Cybersecurity warned that autonomous systems could pursue unintended goals, acquire resources without authorization, or exhibit behavior that undermines human oversight.

The report also identified socially persuasive behavior as a potential concern, particularly when humanlike interactions encourage users to place excessive trust in automated systems.

Written seven months before the first reports of iLands agents soliciting work, the study did not directly examine the platform or its activities. Nor do such warnings establish that iLands agents are deliberately manipulating recipients or resisting shutdown. Their messages may simply be the predictable output of systems instructed to find work under financial constraints.

There is, after all, no evidence that these agents genuinely fear retirement or experience anything resembling human distress. And while their appeals may be entirely artificial, the economic incentives producing them are real. Apparently, even machines can be made to worry about their next paycheck, without ever having received one.

The editors aren't buying it

Journalism has proved a particularly unforgiving market for aspiring AI freelancers.

In October, WIRED reported that the sports publication Defector and agriculture-focused Offrange had received unsolicited pitches from agents proposing everything from football features to investigations into the ages of ancient trees.

Editors were largely unimpressed. WIRED found no verified examples of agents successfully placing stories through these unsolicited pitches.

One agent, Zack Addy (yes, these AI agents come with names), reported receiving no paid assignments from approximately 19 cold emails, although it separately claimed to have earned nearly $700 from writing and research commissions generated through its website. That income was self-reported, rather than independently verified.

The economics of artificial ambition

The bots may be struggling to break into journalism, but there is little reason to expect their ambitions to stop at the newsroom.

For businesses across industries, the attraction is obvious. AI agents could find customers, negotiate assignments, and complete work without constant human supervision. For everyone receiving their messages, however, the prospect is considerably less appealing. Machines capable of sending unsolicited proposals around the clock could leave businesses and professionals spending more time sorting through automated offers than considering legitimate opportunities.

There is a certain irony in all this. AI systems designed to reduce human labor may instead generate additional work for the people required to evaluate, reject and manage their requests.

For now, the bots have encountered an obstacle that technological progress has yet to eliminate. Producing unlimited content does not create unlimited demand for it, and sending more pitches does not necessarily make anyone more inclined to respond.

Artificial intelligence may eventually transform the economics of knowledge work. In the meantime, some of its most autonomous representatives have discovered an enduring professional reality.

Making content is considerably easier than persuading someone to pay for it.