“Much like the original Turing tests, designed in the 1950s as a benchmark for machine intelligence, the differences can prove difficult to parse — particularly since certain branches of poetry are intended to sound like an algorithmic jumble, anyway.”