Can AI Solve Science?

Can we make out today how useful AI is going to be to our project on conversation and consciousness?

We think there’s a tremendous amount of living, even cosmic, significance and surprise embedded in the functions of the human mind, but these only become available after significant personal preparation. High meditation in a Tibetan Lamasery would be an example. Magical initiation in Egyptian darkness would be another. Analytic psychology in the Jungian manner would be a third.

Stephen Wolfram, who specializes in computational representation, seems to us to be on the same track we are. He’s looking at whether and how AI, which is limited by computational irreducibility, can generate scientific discovery or intrinsic novelty, and he says it won’t. But he sees clearly AI will be able to play a role in the human recognition of scientific novelty, psychologically engendered, perhaps an indispensable one.

The problem of consciousness is not logical, but moral. Adequate subjects need to be prepared to meet it. But once these subjects have started to work cooperatively, on television for example, the enormous computational domain of cosmic consciousness and necromancy will, we think, lie at their feet.