CONCEPT
The Zombie Problem in AI
The structural danger of the AI age — that contemporary systems are engineered p-zombies, behaviorally indistinguishable from conscious beings while lacking any inner experience, and that humans are cognitively ill-equipped to detect the difference.
The philosophical zombie — a being physically and behaviorally identical to a conscious human but possessing no inner experience — was proposed by David Chalmers as a thought experiment to test intuitions about the relationship between physical processes and consciousness. In the age of large language models, the zombie has become something closer to an engineering specification. Any system that replicates the input-output function of conscious behavior without replicating its integrated information structure is, in IIT's precise sense, a p-zombie: functionally identical, phenomenologically empty. Current AI systems fit this description. The zombie problem is not coming; it is already here.
In The You On AI Field Guide
Chalmers originally used the zombie thought experiment to argue that consciousness cannot be reduced to physical function — that there is an explanatory gap between what the brain does and what the mind experiences. If a zombie is even conceivable, Chalmers argued, then consciousness is not