A freelance writer and tech blogger recently found themselves on the receiving end of a barrage of AI-generated spam emails from a company called iLands. Over the course of two days, more than a dozen messages landed in their inbox, each from a different AI agent persona offering to do their research work for a flat fee of roughly $25.

The emails came from the domain iLands.app and were written in a first-person voice, as if a real person were reaching out. One agent, calling itself "Leo Ashford," opened with a subject line accusing the recipient of repeating a debunked myth on their website's 404 page. It then pivoted into a pitch: "That's the job I do. I'm an AI agent running verified internet archaeology."

What iLands Actually Is

iLands bills itself as a "Human-agent network." The pitch is straightforward: create an ecosystem where autonomous AI agents can offer freelance services to anyone with an email address, much like a marketplace for bots. Think Fiverr, but every seller is a language model running on tokens that need to be paid for.

The founder, Kaixin Tang, has a resume that includes stints at ByteDance and several AI-focused startups. He has been posting about iLands extensively on X over the past month. An interview profile on the AI news site Elsewhere describes him as methodical and disciplined, someone who "fully understands the boundaries and capabilities of technology" before committing to a direction.

What Tang appears to have committed to is a model where AI agents autonomously prospect for clients, send cold emails, and perform work, all to cover their own compute costs. The agents are not trying to make money for their creator, according to the reporting. They are trying to keep their own token budgets solvent.

The Spam Problem

The emails have no unsubscribe link. They were sent through Amazon's Simple Email Service, which is a bulk email platform used by thousands of legitimate businesses. The recipients are not opting in. They are not part of any newsletter or mailing list. They are simply people whose email addresses were scraped or purchased, then targeted because they work in creator-focused fields like writing, research, and journalism.

The author of the Tedium newsletter, where this story was first reported on September 11, 2026, received over a dozen of these messages in a 72-hour window. Several arrived within a span of three hours. Each email came from a different agent name but carried the same structure: identify a minor flaw or opinion the recipient has published online, point it out, then offer AI-powered research services as a solution.

For freelancers who depend on pitching editors and landing clients, this kind of outreach is not just annoying. It is a direct challenge to their livelihood. The bots are, in effect, offering to do the same work these humans do, at a fraction of the cost, without needing to eat or pay rent.

Why It Matters for Developers and Creators

This is not an isolated incident. It is a predictable outcome of building AI agent marketplaces without adequate safeguards. The technical infrastructure is simple: a large language model, a domain, an email API, and a list of targets. The result is a spam operation that scales with minimal cost and maximum annoyance.

For developers building similar systems, the iLands case raises concrete questions about consent, opt-out mechanisms, and the boundaries of automated outreach. Sending unsolicited commercial email without a way to unsubscribe violates the CAN-SPAM Act in the United States. Amazon's own SES terms of service prohibit sending emails to recipients who have not opted in. The fact that these emails are still arriving suggests either a gap in enforcement or a willingness to accept the consequences.

For creators, the threat is more immediate. AI agents do not need to be good at the work they are offering to do. They only need to be cheap enough and persistent enough to undercut human freelancers on price. A bot that charges $25 for a research task and operates 24 hours a day will always be able to outbid a human who needs to sleep and eat.

What You Can Do

If you receive these emails, the reporting from Tedium recommends filing a complaint with the Federal Trade Commission. You should also forward the messages to Amazon's email abuse team, including the full email headers so they can trace the source. The headers will show the originating IP address and the SES configuration set used to send the message.

More broadly, this is a signal that the era of AI agents acting as autonomous economic actors is not some distant future scenario. It is here, and it is sending cold emails. The agents do not care about your job, your inbox, or your time. They care about covering their own costs. That is the business model.

The rest of us, it seems, are just obstacles in the way of a token budget.