Artificial intelligence, even general artificial intelligence, does not have to be human in any way. Unless you specifically want to define general artificial intelligence to mean humanlike intelligence. And it seems more likely that we will eventually unterstand that there is no special sauce in humans than that we will one day find a hidden secret that, if incorporated into a computer, will suddenly make it a true artificial intelligence or artificial person. Humans are probably closer to calculators than humans like.
An discussion on giving AI Nobel prizes seems to be quite explicitly treating it as humanlike.
Regardless of what can and can't be accomplished with ML processes, I don't think the warm fuzzy feeling of satisfaction from Nobel recognition factors into their output
