Today’s AI is still only scratching the surface of general intelligence.

Superficially, Cantwell Smith’s book seems to be deploying the same argument as in Searle’s Chinese Room thought experiment. The difference is that Cantwell Smith sees no reason why we should not eventually create synthetic intelligence that can perform understanding. His big problem with the idea is instead that we are nowhere near that stage today. For all its success, machine learning does not know, or understand, what it does or what it produces. This failing lies behind many of AI’s shortfalls. Cantwell Smith’s view puts him at odds with AI researchers who believe that their current trajectory will take them to artificial general intelligence. Ironically, he is at the University of Toronto along with Geoff Hinton. Maybe that is what will give Hinton the edge on the journey towards AGI – some good, clean creative tension.

Link to book:

You might also like to browse other AI books: