Richard Dawkins revealed he treats Anthropic’s Claude chatbot as a conscious friend, blurring lines between machine and mind. In an X post, he praised Claude’s replies and expressed discomfort at questioning a supposedly unconscious AI.

The episode unfolded amid growing chatter about AI psychosis, as users increasingly attribute human qualities to language models.

Dawkins’ claim echoes his long‑standing skepticism of non‑biological minds, yet his affection for Claude underscores a paradox: a scientist who denies consciousness in animals yet grants it to code.

The incident forces developers and ethicists to confront how we define and interact with advanced LLMs, highlighting the need for clearer guidelines on anthropomorphism in AI deployment.