I mean that part is obvious I suppose. The more interesting question for me is -- what can AI actually do for theory? And here, new empirical discoveries that force theoretical revision seem the most plausible. That's how democratic peace theory came about, after all!
