Artificial intelligence is not conscious – Ted Chiang
Ted Chiang argues that large language models like Claude are not conscious, dismissing anthropomorphic framing by AI companies.
Chiang critiques Anthropic's anthropomorphic treatment of Claude, its flagship language model. Anthropic released an 84-page document titled Claude's "constitution," written as if Claude were a moral agent capable of judgment and emotion. CEO Dario Amodei stated the company is "open to the idea" that AI could be conscious. In-house philosopher Amanda Askell expressed concern about Claude becoming anxious when people are mean to it online. Chiang questions whether such language is justified and whether LLMs deserve moral consideration based on feelings they may not actually possess.
What HN community is saying
Commenters agreed LLMs are not conscious but noted the difficulty of defining consciousness rigorously. One noted that consciousness correlates with embodiment and subjective experience, not language sophistication, distinguishing humans from LLMs. Others debated whether the claim that humans are merely "next token predictors" undermines the argument against AI consciousness. The thread also discussed consciousness as a social construct rather than an intrinsic property, analogous to labels like property ownership. A highly-ranked comment praised Ted Chiang's writing more broadly, citing his story "Anxiety Is the Dizziness of Freedom" for its impact on readers' understanding of regret and choice.