a dead man carrying his own corpse
A startling image from Mohammed el-Fayturi's poem "Yaaqut al-Arsh." al-Arsh was a Sufi philosopher who died in 1287 in Alexandria, at the age of eighty. Our poet el-Fayturi, who spent much of his youth in Alexandria, died in 2015 at the age of eighty-five. The poem is from Modern Sudanese Poetry, a nice anthology. It will be hard not to steal the image of a dead man carrying his own corpse to his place of dying. A week from now, I will have forgotten all of the poem except this image, and will be convinced that it's my own invention.
Tangentially, the "artificial intelligence" program ChatGPT came up in a work conversation recently, when a colleague said that he'd asked the program to write three sample personal statements suitable for university applications. All three personal statements were similar to each other, and also similar to actual personal statements written by human university applicants. My colleague predicted that AI programs would be writing almost all of these application essays in the near future, and he hopes that this will end the practice of personal statements in admissions packages. I find all of this amusing.
I also found it interesting enough that I went to the ChatGPT website, created an account, and asked the program to write a poem about a dead man carrying his own corpse. I asked for three poems. All of them were awful. Really, truly bad poems. I know that a lot of people are being wowed by AI-produced graphics (which rightfully worries a lot of graphic designers and illustrators), but I think that poets are safe, for now anyway. I also asked the program to write a couple of short stories, which were equally bad for a variety of reasons, and I come away from the experience with some disappointment. Not that I want computer programs to be able to do creative writing, but I think I'd been led to believe that they already could do it, when in fact they cannot. Which is a good thing, despite my adolescent excitement about science-fictional futures. When a computer can actually surprise itself, we'll be in some trouble, lads.
Pretty much my conclusion as regards AI and poetry... A million metal monkeys won't re-invent Shakespeare.
ReplyDeleteI do think it likely that many self-publishing writers will use AI images. And these companies that just reprint domain works with covers that don't match up will as well.
The AI thing seems pretty overblown, except for instances when the text-generator has been given samples of fairly common texts that it can mimic. I do not think there is any art being generated by computers. The AI images are almost--I say almost, but no more--interesting, but they seem by and large to be collages of other people's work, assembled loosely according to keywords. There is of course no creative mind at work, no intentionality, and any beauty or even simple visual interest is purely accidental. I had one of the online programs generate an image of Darth Vader ballet dancing, and it is amusing, but a million or so people could've produced something much better.
ReplyDeleteThe audience for computer-generated art is, I think, simply the audience for mass culture, not people interested in art. Artists will of course be following developments in AI graphics, but not looking for pointers to the way forward as creative humans. I'm feeling very pessimistic today. In about 2005, John Berger wrote an essay in which he talked about how the global mass consumer market and the culture of buying as a way of belonging to the world has replaced actual human culture; what we buy, and the fact that we interact with the market, has become the idea of home, and who we are, and where we are going. Actual creative work, that examines the lived experience and possibilities of humanity, no longer matters unless it can be endless replicated and turned into property that can be sold over and over. Culture-free decorative products, yes. Art, no.
I'm in my cranky old man mode today, I see.
Alas, I do not disagree. Cranky young crone here!
Delete