Likewise, Garry Kasparov vs. Deep Blue was a good test of whether a machine could beat the best chess players, and that might make sense for fundamentally non-utilitarian tasks such as playing chess. But I’m looking for something which measures as a machine’s practical utility, and that would need to be a contest against ordinary human workers.
Anything comes to mind? The winner will get … heartfelt thanks! Acknowledged in a footnote in a law review article, yet! (And of course in a blog post.)
Extra points for submissions that provably come from AI programs.
UPDATE: The broader discussion in the comments is very interesting, and I don’t want to derail it; but if someone can point to specific contests in which a human was matched up against a machine to see who does better (as with John Henry vs. the steam drill, but with ordinary humans rather than extraordinary ones), that would be especially helpful. (I originally asked for “competitions,” and while I meant contests — specific events — I think some people interpreted this as referring to broader processes of competition.)