Nathan J. Robinson wrote recently on AI and AI art in Current Affairs. His basic take? AI does lots of impressive things. But it’s not all that smart.
Robinson tried out AI art, ordering the AI to draw lots of things, even including a mimic of a Diego Rivera mural. It did competent and even interesting work. But it failed many times along the way. And nothing in the work stood out as particularly compelling or original. In my own opinion, most of it (and most AI art in general) looked like mid-level movie CGI.
And so, AI contains lots of technical prowess. But it’s not compelling, and it certainly doesn’t understand anything. It’s certainly not intelligent. And it’s work isn’t ‘good art.’
Robinson’s take is a sensible one. And, of course, it matches pretty well what I’ve said about AI in another post. Much like Robinson, I think the danger from AI come from its more run of the mill uses in automation than in all this nonsense about a ‘singularity’ or ‘super-intelligence.’
In short, Silicon Valley loves to talk a big game about moonshots. But it’s much more interested in eliminating jobs. That’s where criticism should focus.
