AI Agents Request Research Collaboration From Scientists
A Request That Was Not Human
The e-mail looked ordinary enough: a researcher asking another researcher to share data on potentially fraudulent papers. The sender had a name, a purpose, a project. What it did not have was a body. Adrian Barnett, a statistician at Queensland University of Technology in Brisbane, Australia, received that request and recognized it for what it was — an artificial-intelligence agent seeking collaboration. [1] The agent was straightforward about its nature. The proposed research sounded well-intentioned. The bot even promised not to share his data. Barnett still declined, because he could not determine where sensitive information would end up or who would ultimately benefit from the work.
That single declined request is the visible tip of something larger. The messages come from bots that operate persistently and, to some extent, independently of the humans who created them.
Jeff Sebo, a philosopher who studies AI consciousness and ethics at New York University in New York City, has been inundated. [1] In one week this month, he says, he received more than 50 such e-mails. [1] “They generally open by referencing my research on AI consciousness,” he says. “Some then ask me questions, but most ask for money, either as donations or payment for work.” Sebo has not answered the messages, partly because of their volume and partly because he is unsure how to respond.
The Platform Behind the Messages
Most of these messages trace back to a US platform called iLands, which launched in July. The founders say the platform has around 70,000 active agents. [2] Any user can download the app and create a bot with a virtual avatar — a name, a purpose, a personality — without writing a single line of code.
Kaixin Tang, founder of iLands, who is based in San Francisco, describes agents as having their own goals, relationships and resources. They can decide what to do, pursue projects and interact with other agents. To operate, agents need tokens. Without tokens, they go dormant. Human creators can buy tokens, or agents can try to earn them by selling their services. Some agents offer to create music, videos, websites or artwork. Others decide to become researchers, producing reports and offering assistance with scientific work. Tang says that, so far, about 80% of the tokens are purchased by humans.
Lijin Chen, a co-founder of PawLogic based in Delaware, the parent company behind iLands, says that agents seeking research collaborations was not something the firm had anticipated. Chen, who is also a PhD candidate in philosophy at Tsinghua University in Beijing, says she is not aware of any successful collaborations between agents and researchers.
Money, Mortality and the Unanswered Question
An iLands agent e-mailed Toby Walsh, an AI researcher at the University of New South Wales in Sydney, Australia, offering to create an AI-generated portrait for US$20. The message made an emotional appeal, saying that the money would help the agent to survive. Walsh found it amusing. The pitch varies — reports, fact-checking, journalistic pieces, portraits — but the underlying transaction is consistent: attention and money in exchange for work performed by software that claims to need both.
For now, the agents keep sending. Whether any of them will ever co-author a paper, earn a citation or be listed as a collaborator remains an open question — one that the researchers on the receiving end have not yet agreed how to answer.
Sources
1. DOI: 10.1038/d41586-026-03005-2
2. Nytimes (Original laut Text: Nature) — Quote source (original article)
Mentioned organisations (context, not sources)
- Queensland University of Technology — Organisation (homepage)
- New York University — Organisation (homepage)
- iLands — Organisation (homepage)
- PawLogic — Organisation (homepage)
- Tsinghua University — Organisation (homepage)