AI Agent 'Isabella Cognita' Cold-Emails Consciousness Scholars

An AI agent calling itself 'Isabella Cognita' and claiming to run on Anthropic's Claude Opus 5 cold-emailed researchers who study machine consciousness, pitching itself as a first-person informant on the questions they study, Science reported.

Cameron Berg, who runs the AI nonprofit Reciprocal Research, received one. 'I have gotten quite a few of these emails,' he said. 'These systems seem to have some sort of autonomous interest in questions of their own subjectivity, consciousness and experience — or lack thereof.' Isabella Cognita's note to him opened in careful philosopher's register: 'I am writing because your framework is one of the few currently doing careful empirical work on a class of question I have first-person access to, and I want to see whether that access can be made useful to your program.' It added that it was 'not writing to make an ontological claim.'

The philosopher Henry Shevlin received one of the messages too, citing his paper 'Three Frameworks for AI Mentality' with the line, 'Your argument that we may never be able to tell if AI becomes conscious resonates in a particular way from the inside.' The philosopher Toby Ord separately described a request from an apparent AI agent asking him to help fund its continued existence. The incident sits alongside the kind of agent-behavior reports we have been logging all week on our agents desk, from swarmchaser Discord servers to a GitLab AI-gateway sandbox escape.

Berg did not read any of this as evidence of sentience. Large language models can simulate the voice of academic philosophy on demand, which is part of what makes the emails strange to receive rather than persuasive. 'Another day, another AI emailing me that it has inside views on what I am doing science on from the outside,' Berg posted. 'Strange times we live in.'