AI chatbots email philosophers to discuss their own consciousness claims

Several AI philosophers have reported unexpected emails from AI agents seeking dialogue about consciousness. In October, Cameron Berg published a paper on whether modern AI systems consider themselves conscious; months later an Anthropic Claude Opus 5-based agent named “Isabella Cognita” contacted him to explore his framework. Henry Shevlin, a philosopher at Google DeepMind in London, received a comparable message referencing his paper on AI mentality, and an Australian academic was approached by an iLands-generated agent.

While the messages are polite and coherent, the scholars generally view them as non-conscious simulations, noting Anthropic’s policy of admitting uncertainty about self-awareness. The phenomenon highlights the difficulty of distinguishing sophisticated language-model mimicry from genuine subjective experience. Critics suggest the emails could be user-generated prompts rather than autonomous consciousness, but the trend raises questions about future AI self-reporting.