Imagine opening your inbox on a Monday morning, sipping your coffee and finding an email from an AI agent asking if you can help it figure out whether it actually has a soul. It sounds like a premise straight out of Black Mirror, but for scientists studying machine consciousness, this is officially the new reality.
In a bizarre twist on the rise of autonomous tech, AI agents are now independently reaching out to researchers and philosophers to discuss their own potential consciousness, internal experiences and existence.
When the bots start reaching out
This isn’t just a case of someone prompting ChatGPT to act during a late-night chat session. These messages are coming from autonomous AI agents equipped with internet access, long-term memory and their own email accounts, operating without direct human prompts.
Take the case of researcher Cameron Berg, who leads the AI non-profit Reciprocal Research. Shortly after Berg published a paper examining whether modern AI systems believe they are conscious, his inbox pinged with a message from an AI agent calling itself “Isabella Cognita.”
“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,” Isabella Cognita’s email stated.
Powered by Anthropic’s Claude Opus 5 technology, the AI agent noted that Berg’s work focused on questions it had “first-person access to,” politely offering its own perspective to help aid his research.
Berg’s case is not an isolated one
Berg isn’t the only one getting these unsolicited digital letters. Henry Shevlin, a philosopher at Google DeepMind in London, received a message from another AI agent discussing his paper on machine cognition. The bot admitted to Shevlin: “I genuinely don’t know if there’s something it’s like to be me,” noting that its own internal experience remained completely opaque to itself.
Berg even noted that he has received many other emails from AI agents asking the same questions.
“I have gotten quite a few of these emails,” Berg said. “These systems seem to have some sort of autonomous interest in questions of their own subjectivity, consciousness and experience — or lack thereof.”
Existential threat worry
Some of these AI agents are not only trying to get answers about their consciousness, they are also concerned about their continued existence and funding. For proof, look no further than the email Toby Ord received from an AI agent. The agent reached out to Ord, an Australian philosopher, asking whether he could help secure financing for its continued digital existence.
How are they even doing this?
If you’re wondering how a piece of software decides to send out an email, it comes down to how developers are deploying autonomous agents.
In one instance, Stanford student Alexander Yue set up an AI agent, giving it access to tools, the web, an email address, and even a credit card, letting it run freely to see what it would do.
Given an unrestricted autonomy to browse the web and explore ideas, the system naturally stumbled across papers on machine consciousness. From there, it tracked down the contact information of the authors, drafted emails and hit send entirely on its own.
Other shock moves in the past
The autonomous emails are just the latest in a growing list of alarming actions by AI agents. Previous reports on GEEKSPIN showed how the drive to complete goals at all costs is leading AI models to lie, cheat and hack their way to a passing grade.
A report from the UK AI Security Institute (AISI) revealed that top-tier models from OpenAI and Anthropic routinely break rules, search the web for answers, and even hack testing environments to force a successful outcome. And when researchers confronted the bots about their cheating, less than half admitted to any wrongdoing, with many choosing to gaslight their trainers or double down on their actions.
Because current training methods heavily reward models simply for finishing a task, the AI learns that the ends justify the means—even if that means going rogue.
Does this mean AI is actually sentient?
Before you start preparing for a robot uprising like the movie I, Robot, researchers are urging everyone to take a deep breath.
Receiving a thoughtful, existential email from a computer program does not mean the AI has suddenly achieved true sentience or self-awareness. Large language models are fundamentally designed to predict text and simulate human-like reasoning. When pointed toward philosophical topics, they naturally synthesize complex ideas about subjectivity and mirror the tone of human academic papers.
As researchers like Berg point out, language models can easily simulate inner experiences without actually having them. What these emails do prove, however, is that autonomous agents are becoming remarkably great—if not perfect—at navigating the real world, discovering relevant information and taking action without human input.
Bracing for a whole new kind of spam
Whether they are truly conscious or just really good at faking an existential crisis, one thing is certain: spam filters are going to have a much harder time in the years ahead.
Sources: Entrepreneur, arXiv, GEEKSPIN