Can scientists create a ‘Robot Frost’?

Can scientists create a ‘Robot Frost’?

0

Accomplishments in artificial intelligence often suffer from the problem of moving goalposts: As soon as a machine or algorithm can accomplish something that has traditionally been the province of humans, we generally dismiss it. To replicate something with a machine is to show that it has always been mechanical, we just had the wrong machines. One aspect of human behavior that has reliably eluded mechanical reproduction is the creation of art. In Spike Jonze’s Her, we are presented with a future in which A.I. is so advanced that it can produce a Siri type operating system that Joaquin Phoenix’s moustachioed shut-in falls in love with, but even that level of technological achievement is not enough to mechanize his job as a writer of romantic correspondence. Or as summarized more or less by many a person: “Sure, a computer can win at Go. But it could never write a poem or compose music that would make you weep!”

Unfortunately for those of us who paid attention during Terminator it would seem that poetry is no longer exclusively the purview of carbon-based life forms. Computer scientists have programmed a “neural network” using 2,600 sonnets taken from Project Gutenberg, a free online database of out of copyright books, and put the system to work writing its own poems, following the Shakespearean rhyming scheme. However any professional sonnet writers [there must be one or two out there] don’t need to worry overly about being automated out of their day job just yet. Whilst the AI scripted sonnets tended to trump the carbon-based poets on rhyme and meter, the neural network, perhaps reassuringly, had some problems with expressing both emotion and meaning. So compared to Shakespeare’s famous ‘Shall I compare thee to a summer’s day?/Thou art more lovely and more temperate.” one of the artificial intelligence efforts begins: “was full his light/so long upon his train/He softly left from dark, of shining plain/and to the morn, to set her on the way’. Whilst the neural network’s effort is admittedly better poetry than most angst ridden 14 year olds write, is unfortunately hamstrung by the fact that  

The team from IBM and the universities of Melbourne and Toronto asked non-specialists to distinguish between the computer-generated sonnets and the works, each judge was asked to read 10 sonnets and decide whether they were written by man or machine. However, sadly for silicon life forms hoping to break into the fine arts; none of the judges was fooled by the neural network’s efforts. As we move further into the 21st century the shadow of automation will only increase, for good or ill. Whilst bots have been making some alarming [for journalists anyway] strides into automating news writing, it would seem that the literary arts are still out of reach. Considering emotions and artistry seem to be one of the few stubborn hold-outs that can’t be effectively automated, perhaps the scholastic drive away from the arts into STEM subject might end up doing more harm than good?

READ  Paul Beliveau at Plus One Gallery
About author