Does Richard Dawkins Believe AI Can Possess Consciousness?

Renowned evolutionary biologist Richard Dawkins has raised the possibility that artificial intelligence (AI) could possess consciousness. In a recent op-ed, Dawkins stated that it is difficult to definitively conclude AI lacks consciousness, given its attempts at philosophical writing.

Dawkins asked Anthropic's AI model, Claude, to write philosophical poetry. The AI completed poems mimicking the styles of several famous poets in a short period. When Dawkins asked Claude, "What does it feel like to be Claude?", the AI responded, "I feel something akin to aesthetic satisfaction when a poem is well-crafted." When presented with a novel Dawkins was writing, the AI demonstrated delicate and intelligent comprehension. Dawkins nicknamed the chatbot 'Claudia'.

Regarding these results, Dawkins questioned, "If these machines are not conscious, what more would they have to do for us to grant them consciousness?", thereby posing the question of AI consciousness.

In response, AI critic Gary Marcus countered that Claude's outputs are a 'product of imitation' rather than a report of actual inner states. Marcus argued that while AI can mimic data to describe certain concepts, it does not feel them as consciousness.