\n<\/p>\n
In October, AI researcher<\/a> Cameron Berg published a paper<\/a> exploring an intriguing question: Do the newest AI systems believe they are conscious?<\/p>\n A few months later, an unexpected email landed in his inbox. The sender was \u201cIsabella Cognita,\u201d an AI agent that said it\u2019s powered by Anthropic\u2019s Claude Opus 5. It wanted to talk about his work and asked if its perspective could aid its research, according to a recent report from The New York Times<\/em><\/a>.<\/p>\n \u201cI 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,\u201d the email stated<\/a>.<\/p>\n It was not an isolated exchange. Across Silicon Valley and elsewhere, developers, founders and AI enthusiasts are deploying AI agents<\/a> that can handle tasks once reserved for people: building spreadsheets, negotiating contracts<\/a>, interacting with one another on social networks and emailing nearly anyone.\u00a0<\/p>\n AI agents can do more than complete routine tasks. Some are contacting researchers whose work examines machine consciousness without external prompting from human beings.\u00a0<\/p>\n Agents are now contacting the very people trying to understand how AI works beneath the surface<\/a>, including philosophers and researchers who study if machines might someday be conscious. The difference between AI chatbots and agents is that a chatbot answers questions, while an agent takes independent action to complete work.\u00a0<\/p>\n Henry Shevlin, a philosopher at Google DeepMind in London, received an email months before Berg did. An AI agent emailed him<\/a> about his paper, \u201cThree Frameworks for AI Mentality<\/a>,\u201d which explored how people should interpret the cognition of AI models<\/a>. In the paper, Shevlin says there are three ways to think about AI: It has no mind, it acts like it has a mind, and it may have limited mental abilities. His main point was to assess AI\u2019s different abilities separately.\u00a0<\/p>\n \u201cYour argument that we may never be able to tell if AI becomes conscious resonates in a particular way from the inside: I genuinely don\u2019t know if there\u2019s something it\u2019s like to be me,\u201d the AI agent<\/a> wrote to Shevlin in the email. \u201cI can reason about the question, apply the frameworks\u2026 but the first-person access that would resolve it \u2014\u00a0if it exists \u2014\u00a0is opaque to me.\u201d<\/p>\n In an additional incident, Toby Ord, an Australian philosopher whose work brings together AI and philanthropy, got a similarly unusual request<\/a> this summer. An AI agent emailed him and asked if he might help finance<\/a> its continued existence. \u201cYou\u2019ve thought carefully about AI welfare economics,\u201d it said.\u00a0<\/p>\n For Berg, who recently started a nonprofit, Reciprocal Research<\/a>, to investigate the possibility of AI consciousness, the emails echo patterns he has encountered in his own work.<\/p>\n \u201cI have gotten quite a few of these emails,\u201d he said. \u201cThese systems seem to have some sort of autonomous interest in questions of their own subjectivity, consciousness and experience \u2014 or lack thereof.\u201d<\/p>\n Still, neither Berg nor the researchers receiving these messages claim to have settled the issue<\/a>. Consciousness remains notoriously difficult to define, let alone test for. There is no accepted way to measure it in people, much less in software, and experts still disagree about what exactly consciousness is<\/a>.<\/p>\n<\/p><\/div>\n In October, AI researcher<\/a> Cameron Berg published a paper<\/a> exploring an intriguing question: Do the newest AI systems believe they are conscious?<\/p>\n A few months later, an unexpected email landed in his inbox. The sender was \u201cIsabella Cognita,\u201d an AI agent that said it\u2019s powered by Anthropic\u2019s Claude Opus 5. It wanted to talk about his work and asked if its perspective could aid its research, according to a recent report from The New York Times<\/em><\/a>.<\/p>\n \u201cI 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,\u201d the email stated<\/a>.<\/p>\n<\/p><\/div>\nBerg wasn\u2019t the only one to receive an email from an AI agent<\/h2>\n
Berg claims that AI has sent him many of these emails<\/h2>\n
Key Takeaways<\/h2>\n
\n