The bookâs core argument is that an artificial intelligence that could equal or exceed human intelligence â sometimes called artificial general intelligence (AGI) â is for mathematical reasons impossible. It offers two specific reasons for this claim.