Sla­te just published an arti­cle on Eure­ka, a machi­ne that attempted—and some­what exceeded—to auto­ma­te the pro­duc­tion of poe­try. Rather than being a curio and a side­show of the time, Eure­ka was ins­te­ad, “The Vic­to­ri­ans loved a show, espe­ci­al­ly one that drew upon both sci­ence and illu­si­on for an unf­or­gett­able per­for­mance. The Eure­ka offe­red both. The machi­ne was a feat of indus­tri­al pro­gress, but sen­sa­tio­nal enough to be enter­tai­ning. And it can still wow. The unvei­ling of a recent­ly res­to­red ver­si­on of Clark’s inven­ti­on drew crowds of exci­ted visi­tors who, like spec­ta­tors past, reveled in the show.”

But can a machi­ne ever real­ly pro­du­ce poe­try? Well, an inte­res­ting take on it was pre­sen­ted by Oscar Schwartz, at a TED Talks event in Syd­ney, Aus­tra­lia. Schwartz, who co-crea­ted the Bot or Not web­site, said that any poe­try pro­du­ced through machi­nes reflec­ted the input—that if you put Emi­ly Dick­in­son into a machi­ne and asked it to pro­du­ce a poem, it would pro­du­ce an Emi­ly Dick­in­son-like poem. Qui­te sim­ply, you get out what you have put in. This, says Schwartz, rai­ses inte­res­ting ques­ti­ons about how we defi­ne what it is to be human.

Ulti­mate­ly, poe­try-gene­ra­ting machi­nes fail. Not becau­se they fail to pro­du­ce poems but becau­se they fail to pro­du­ce any­thing out­side the nar­row lines it is given. We can teach a machi­ne to do some­thing, but they’re not yet at the point whe­re the machi­ne can learn it by them­sel­ves.

But perhaps the most-tel­ling line of the Sla­te pie­ce is its first one: “Poe­try allows us to explo­re the dep­ths of what [it] means to be human.” A poem from a machi­ne can­not reso­na­te like Leo­nard Cohen wri­ting to a dying love; con­nect us like Ray­mond Carver’s ‘Late Frag­ment’; draw in poli­ti­cal and soci­al con­text as Muham­mad Ali—probably inadvertently—with, “Me./We.”; or rai­se a wry gigg­le like Roger McGough.