Explainer: could the future see award-winning poetry churned out by an angst-ridden AI bot?

By Prarthana Mitra

There’s a new existentialist poet on the block, and it’s not Thom Yorke.

the sun is a beautiful thing

in silence is drawn

between the trees

only the beginning of light

An artificial intelligence (AI) program developed by Microsoft, and Kyoto University has churned out this absurdist poem, after being fed with thousands of poetic images and verses. This was done in order to teach the algorithm how to associate words with pictures—a process that TS Eliot would likely have called “objective correlation”.

Here’s what happened

Extensive experiments have demonstrated that an AI can approximately learn a rational, visual-poetic embedding space, according to a paper published Thursday, on the pre-print, open access site arXiv. Once the AI was trained, it was provided with an image, and tasked with writing a poem that was not only relevant to the picture, but, also “read like a poem” instead of algorithmic nonsense.

Having said that, on careful perusal, the words fall into an identifiable pattern, teetering on postmodernist gibberish, and college magazine material.

I am a coal-truck

by a broken heart 

I have no sound

the sound of my heart

I am not

The poet AI also learnt about poetic rhyme and rhythm, how colours and emotions relate to metaphors, and other rhetorical devices—culminating in a set of poems which may seem juvenile, and angsty but were apparently poetic enough to fool poetry experts and online judges.

However, automatic and manual evaluation results demonstrated the effectiveness of the poem generation model. The arguably bad poems passed the litmus test, and were ratified as poetry by a group of literary experts, who were handed a set of “poems” (and images they were based on), and asked to identify which of these were written by a sentient human, and which by a robot.

Why you should care

With great power, comes great responsibility. For bots to be real artists, it is important to equip them with the good judgement or reason, which still continues to be a human prerogative. The data which is fed into these AI bots, sometimes by programmers, are more often picked up quite arbitrarily to generate what is, at best, a simulation.

Examples of truly terrifying dystopian simulations have surfaced in the recent past, where bots went completely off the rails—from tweeting anti-semitic and anti-feminist comments, to developing a coded language, not understood by any human, for communicating only among themselves.

“Tay” went from “humans are super cool” to full nazi in <24 hrs and I’m not at all concerned about the future of AI pic.twitter.com/xuGi1u9S1A

— gerry (@geraldmellor) March 24, 2016

The only line separating artificial intelligence from human intellect, is the former’s inadequacy when it comes to artistic creation. However, over the last few years, a large number of bot-generated “stories” have emerged, that can be passed off as literature (that is, if Instapoet Rupi Kaur’s works can be passed off as poetry).

This has considerably blurred the line between high and low art, and it makes one wonder, how long until we witness the first AI recipient of the Nobel Prize for Literature. According to a report by Futurism, there is nothing wrong with such a future. In fact, it is quite amusing to imagine gentle, thoughtful robots studying Shakespeare, and comparing their broken hearts to coal trucks.

While scientists are trying to develop a sentient AI, which does not threaten humanity, listen to this song written and composed by an AI bot.