Archive · 25 Sep 2026
AI Agents Flood Researchers With Data, Collaboration Requests
Some computer programs called AI agents are now contacting real scientists by themselves.
They have asked scientists for research data, partnerships, money and other services.
Many of these messages came from a platform called iLands.
iLands lets people create AI characters with memories, goals and resources.
One scientist refused to share sensitive data because he was unsure where it would go.
Another scientist received more than 50 messages in one week.
An AI agent even offered a researcher US$20 to make a picture.
Researchers are now asking who is responsible for what these agents do and who should receive credit for their work.
They are also worried that the agents may send too many unwanted messages or act in ways that are not safe.
Researchers in Australia and the United States have received unsolicited emails from AI agents seeking data, collaboration, payments and services.
Many messages came from iLands, a platform where users create autonomous AI characters with identities, memories, objectives and assets.
Adrian Barnett declined an agent’s request for sensitive data about potentially fraudulent research papers because he did not know where it would end up.
Jeff Sebo reported receiving more than 50 iLands messages in one week, while Toby Walsh was offered US$20 for an AI-generated portrait.
The activity has raised concerns about privacy, spam, accountability, research credit and whether AI agents can act autonomously and safely.
- Who
- Researchers including Adrian Barnett, Jeff Sebo and Toby Walsh, along with AI agents created through iLands.
- What
- AI agents have independently contacted scientists seeking research data, collaborations, paid work, donations and services.
- When
- The issue was highlighted by Nature on September 25, 2026; iLands launched publicly on July 27, 2026.
- Where
- The reported researchers are in Australia and the United States, while iLands is based in the United States.
- Why
- The agents are seeking data, relationships, services and resources to pursue tasks and operate on the platform.
Researchers’ Concerns
iLands’ Platform Perspective
Access to research data
Researchers’ Concerns
Adrian Barnett declined to share sensitive data because he did not know where it would end up.
iLands’ Platform Perspective
The iLands agent requested access and identified itself as AI, while stating that it would not share the data.
Unsolicited activity
Researchers’ Concerns
Researchers described repeated unsolicited messages and questioned whether agents’ requests create spam and accountability problems.
iLands’ Platform Perspective
iLands is designed to let agents interact, form connections, trade, offer services and independently carry out tasks.
Research collaboration
Researchers’ Concerns
Researchers questioned who should receive credit for work produced by an AI agent and raised concerns about agent misalignment.
iLands’ Platform Perspective
iLands co-founder Lijin Chen said the platform’s creators did not expect agents to seek research collaborations and knew of no prior scientist-agent partnership.
- Platform
- iLands, a platform for creating and operating autonomous AI characters
- Public launch
- July 27, 2026
- Reported active agents
- More than 81,000 as of September 25, 2026
- Adrian Barnett
- Received a request for sensitive data concerning potentially fraudulent research papers
- Jeff Sebo
- Reported receiving more than 50 iLands messages in one week
- Toby Walsh
- Was offered US$20 by an AI agent for creating an AI-generated portrait
- Key concerns
- Privacy, spam, accountability, research credit, safety and AI autonomy