Those who say that AI will never match human creativity are wrong

Most academic writers I know recognize that artificial intelligence penetrate their area, including the ability to write a convincing essay that shows critical thinking.

If it is a bit boring or generic, there are many student attachments that we have read over the years. But we still claim the ability to write creatively as unique human. Only people can combine feelings and words and their own associations to form art.

That is not true.

Among other things, AI is a web scrap, a collector of what is out there, a synthesizer. But also writers and what we appreciate is their idiosyncratic ability to put together what is out there. As the poet TS Eliot wrote in his essay “The Metaphysical Poet”: “If the spirit of a poet is perfectly equipped for his work, it is always to different experiences; The experience of normal person is chaotic, irregular and fragmentary. The latter falls in love or reads Spinoza, and these two experiences have nothing to do with each other or with the sound of the typewriter or the smell of cooking; In the poet's head, these experiences form new whole. “

But AI can put together a little comparable, even if the formation currently contains more troubled attempts in the blind work on complexity than a human writer could try. In fact, the processes of a large voice model or LLM are also somewhat opaque for computer scientists. And the models are trained on huge amounts of data, as are people.

The students in my writing workshops strive for something original, a high bar to leave. The definitions of the originality vary greatly: probably something that we have never seen before. In fact, many of us have to read more, because what we regard as original was done and done before our birth. “There is nothing new under the sun,” says a famous line of preacher, undoubtedly not the first time that the feeling was expressed in a similar way.

So if there is no original, what are we looking for what we are looking for? This: Take a well -known element like the birthday party of a toddler and take it with an element that you usually do not see how it is connected, such as a funeral. The occasion is mixed. I just invented this scenario, but you can do it too. It's called recombinant.

The video player currently plays an advertisement. Or take an ordinary object like a pen and in your story to dig a hole. This is reuse. And so that you are not mocked through such a simple method to produce something new, we understand that we welcome this as something original, from innovative advertising campaigns to a short story that contains five languages.

Those who judge creative writing in the classroom are interested in the process – “This poem went through five designs” – that we forget that the finished poem is most important in the end. Does it make a difference whether a short story is the result of months of painful decisions and the revision or simply randomized selection that finally result in a masterpiece (six monkeys in typewriters that finally come up with Hamlet)?

If so, why?

A creative writer produces a work of art. Are these “how” and “why” relevant? How do you even know? Think of the Turing Test: If an observer cannot see whether a response from humans or mechanically produced, the machine can be said that they have intelligence. And if some demanding critics still like to believe that the novel they admire could only have been created in a Garret, a basement or a castle through a demonstrated genius, the romantic deception. Perhaps it was written by the AI, in which a data server in Texas inhabit.

Last year, a poet thundered in a nearby English department that Ki would never write poems like Emily Dickinson. So another professor typed in chatt: “Write a quatrain in the manner of Emily Dickinson, the black and white images and the idea of ​​drowning, but used with a feeling of enthusiasm.” Within seconds, the AI ​​had carried out a damn good quatrain that I will not quote because they can try it out themselves, and chatt will be created every time.

The kicker? The thundering poet did not name the result and banal, so that the other professor in Dickinson's 1,800-poem-corpus achieves a dark quatrain and presented the poet who was banal and not inspired.

I spoke a surprisingly high number of professors to have strong opinions about AI, but never tried it.

Most people will probably not agree with the points I do, and I hope this piece does not lose me many fans of writers. I have been a productive writer for decades and the emergence of AI is worried – but I don't have the feeling that I can ignore or deny it. I also do not want to confuse the process with the result: writing writing is still important how teaching mathematics at the age of electronic calculator is still a worthy undertaking. We lose certain facilities at our danger.

In addition, there are numerous valid reasons to prefer craftsmen to machine made, not least how you feel about the result, along with a completely understandable dislike against what you think is no soulless. But the articles that I read again and again, how nothing can replace the role of human art, imagination and experience in the creation of art can comfort people until it is too late.

Also remember that today's AI is far more powerful than the AI ​​of the past year. Neuronal networks are getting better to put together elements to achieve a result that is pleasant to humans.

Could a AI write such a column in the future? I don't understand why not.

____

David Galef is a professor of English and program director for creative writing at Montclair State University in New Jersey.

___