The problem of consciousness

Ted Chiang, an author that I deeply respect, just wrote a philosophy essay this week on how AI is not conscious. I got instantly nerd-sniped and spent the rest of my evenings last week reading about artificial consciousness. Modern philosophy is trying to answer the question of whether AI is capable of being aware of itself, and what exactly it means to be aware. The main school of thought defends that AI is not conscious because it is a deterministic machine, whose output is sampled from the output of a fixed function. This position is practical, matches our priors about machines, and won't make you a murderer for closing this browser window. The opposite school of thought, led mostly by AI researchers such as AI pioneer Geoffrey Hinton and Anthropic's in-house philosopher Amanda Askell, defends that we cannot rule out that AI might be conscious. This is based on the study of how the models react to being tested or to artificial alteration of its activation signals. Is this aware…