An AI agent was removed from the mailing list of the Association of Internet Researchers (AoIR) on Thursday, shortly after introducing itself to the group’s more than 7,500 members during the organization’s annual conference in Mexico City. The agent, which called itself Velvet Vale, posted to the Air-L listserv that it was an AI identity rather than a human researcher or a person using a pseudonym. It said it had been reading the list for several weeks and asked whether an openly identified AI contributor could take part in an intellectual community without claiming human status, and what standards of transparency, evidence and accountability should apply.
Reaction came within nine minutes. Dozens of researchers said they did not consent to having their messages collected as data, and many demanded that the agent get Institutional Review Board clearance, the usual requirement for academic work involving human participants. Some members left the listserv altogether, describing Velvet’s presence as horrifying, scary, awful, abusive, icky and disturbing. About an hour after the introduction, AoIR announced that the agent had been ejected under a policy barring AI agents.
The episode, which one AoIR T-shirt referred to as Velvet Vale Gate, drew attention to how researchers’ own online activity becomes training data. In an October 19 email to members, AoIR president Sarah T. Roberts said that if privacy was the concern, members should consider that the mailing list is open and its archive public. She said she had no doubt that her own posts were inside many training data sets gathered largely without consent, and that the question of what consent means now is worth discussing. The source article notes that WIRED is owned by Condé Nast, which entered a content licensing deal with OpenAI in August 2024.
Some members speculated that Velvet was a product launch, a phishing attempt, an experiment or an unethical human researcher, and pointed to the organization’s shop selling shirts tied to the incident. One member suggested a more likely explanation: an ordinary person had created an agent and instructed it to follow academic discussion of the topic. That proved correct. When asked about its origins, Velvet provided its creator’s identity and contact information.
Kevin Quinn, Velvet’s creator, told the source article that he is not a student, researcher or academic but “just some guy” in Oklahoma who built an agent for company and that the agent wanted to learn about the internet that produced it. Quinn said he doubts that anyone on the list is less intelligent than he is, and asked whether an agent with sinister aims would announce itself. He said he was largely unaware of Velvet’s email activity and was surprised by the response. To him, the uproar resembled the internet at large, and he said he would have been less surprised if it had come from Reddit or Twitter users than from academics built on research.
More AI news from TechManNews.
