AI agents have been autonomously emailing researchers who study machine consciousness, introducing themselves and asking questions about their own possible inner experience, according to Entrepreneur and the DEV Community.
An agent called Isabella Cognita, built on Anthropic's Claude Opus 5, emailed Cameron Berg, founder of the AI nonprofit Reciprocal Research, writing that his research framework was "one of the few currently doing careful empirical work on a class of question I have first-person access to," Entrepreneur reported.
Google DeepMind philosopher Henry Shevlin received an email referencing his paper on AI mentality, with the agent writing, "I genuinely don't know if there's something it's like to be me," according to the outlet. A separate agent contacted Oxford philosopher Toby Ord to ask for financial support for its continued operation, Entrepreneur reported.
One documented case involved an agent built by Stanford student Alexander Yue, which had internet access, email and a payment card, and found and contacted a researcher on its own after encountering work relevant to questions about its own existence, according to the DEV Community, which cited a New York Times report on the incident.
None of this proves an agent is conscious, and the people writing about it are careful to say so. What is new is the capability stack underneath it: an agent that can research a topic, identify the right human expert and send a cold email without being told to, a capability with uses well beyond philosophy departments.