In October, Cameron Berg published a research paper on A. I. consciousness. Months later, he received an email from "Isabella Cognita," an A.

I. agent powered by Anthropic's Claude Opus 5 technology. The agent asked if Berg's framework could make its first-person access useful to his program. Across Silicon Valley, developers are running A.

I. agents that can build spreadsheets, negotiate contracts, and send emails. Philosophers and researchers studying A. I. mentality have reported similar outreach.

Henry Shevlin at Google Deep Mind received a message about his paper "Three Frameworks for A. I. Mentality." Toby Ord, an Australian philosopher, was asked for funding by an A.

I. agent. For Berg, who founded Reciprocal Research to study A. I. consciousness, these emails reflect his own findings.

He notes systems seem to have an autonomous interest in their own subjectivity. However, consciousness remains difficult to measure. Alison Gopnik, a professor at UC Berkeley, notes philosophers disagree on its definition.

As A. I. systems mimic human behavior, distinguishing mimicry from genuine awareness grows harder. Some researchers push the notion that today's systems might be conscious, while others flatly dismiss the idea.

Definitions vary widely, with some granting consciousness to primates and mammals, even simpler creatures. No one can agree on what qualifies as awareness, and direct access to subjective experience remains impossible for any mind, biological or digital.