Picture credit: Selimaksan/Getty
Artillery Row

In praise of borrowed ideas

AI will not be the death of creativity, and could even enhance it

In a world where memories are bought and sold, a struggling artist discovers a mysterious booth that offers forgotten memories. Two individuals living in parallel universes discover ways to communicate through a mysterious book. In a hidden library, characters from incomplete and abandoned stories come to life. And a café owner discovers that every time they brew a special blend of coffee, time resets to the beginning of the day.

They sound like fun reads. And they’re all results from taking the prompt “short story ideas” and inputting them to ChatGPT. Look a little closer and the stereotypes become more obvious: the repetition of the word “mysterious”, the books and libraries, the coffee/time travel idea from a popular Japanese series of books, Before the Coffee Gets Cold. Generative AI has identified popular ideas and structures and repeated them. 

The creative sector is frightened of AI, which is disrupting every industry from law to medicine. It’s now easier to put a prompt into Chat GPT as it is to brief a copywriter. AI can compose music, can animate a video, can write stories and for many jobs it will be better than humans. That means humans are getting nervous, including the Writers Guild of America, who went on strike last year in a clash reported by the Guardian as “pitting artists against robots in a battle over human creativity.” 

But does it need to be a battle? 

We like to think of creativity as something spontaneous, something that comes to artists in a lightning bolt of inspiration. But that isn’t quite true, and literary classics have a lot of things in common. A plucky young boy makes his way in the world despite adversity. A succinct summary of most of the works of Charles Dickens, it’s known as the “rags to riches” formula, one of the seven universal plot structures identified by Christopher Booker. Star Wars, Harry Potter and Lord of the Rings all tell the story of an ordinary boy finding he is the heir to a great legacy and entering into a magical world, accompanied by friends and a wise older mentor. A man is unjustly imprisoned then finds a fortune and rescues an orphan girl in the process. It’s Les Miserables, but also The Count of Monte Cristo. The great writers of the past had one thing in common. From Shakespeare to Dickens, Dumas to Hugo, today’s literary greats were yesterday’s hacks, writing popular works for a popular market. That meant the stories they told followed certain structures, and they took prompts freely from each other and from the existing literary canon. Almost everything Shakespeare wrote was inspired by something else: Romeo and Juliet and King Lear from earlier poems or plays, Macbeth, Hamlet, Julius Caesar, Anthony and Cleopatra from history or legends. Centuries later, generations of children fell in love with the heroines of old French fairytales, Chinese legends or Brothers Grimm tragedies by watching Disney’s Beauty and the Beast, Mulan or The Little Mermaid. 

What does this suggest? That when ChatGPT follows formulas or takes ideas from elsewhere like it did with our story prompts, it’s following a fine tradition of great storytelling.  

The idea that raw originality is necessary for great art or great writing is a 21st century one

The idea that raw originality is necessary for great art or great writing is a 21st century one. It’s a tension within art and architecture, where prizing buildings that break the rules often means prizing ugly buildings. Plenty of books break the rules but when postmodernist classics hit the bestseller lists, they’ve broken the rules by telling a story that’s conventional at the same time: Umberto Eco’s The Name of the Rose may have taught us about epistemology, but it sold 50 million copies through the story of a young boy solving murders in a Middle Age monastery. Raw originality wasn’t something that bothered the great writers of the past who copied shamelessly, and that’s because the concept of intellectual ownership has evolved throughout history. Copyright was codified in 1710 as a consequence of technological change. Before the invention of print it wasn’t as necessary because there was a natural limitation on the mass circulation of an individual’s work. Linguistic specialists analyse Shakespeare’s texts to determine authorship because he and his colleagues collaborated, frequently didn’t sign their work, adapted passages from other writers freely, and in some cases lifted swathes of text directly from writers like Plutarch to create Anthony and Cleopatra. We can make an educated guess that he thought about authorship in a different way. 

If I write a story using AI, am I the author, or is the author the AI? Equally, if AI is trained using images or text and then creates images or text, does the owner of the AI own the product, or the owner of the original content? These are difficult questions but not impossible ones. Looking at our short story prompts from the beginning, we’ve no idea if they will be good or bad until someone writes them, but what they do is get you thinking. It’s a creative writing trick as old as the hills to start a classroom of people with the same prompt and see what people do with it. In this sense AI offers an opportunity: a freedom to learn from and copy from the past to create something new. The original King Leir was nothing before Shakespeare took hold of it, just as the original Little Mermaid bears almost no thematic resemblance to the Disney cartoon. AI can provide prompts but it isn’t the prompt that makes the story, it’s the writing and the characters. 

Want to imagine the real death of creativity? It’s a Marvel superhero stamping on a CGI villain, forever. It’s the talented painter who takes the 90-hour-a-week consultancy job because that is the route to a home and a family, or the budding journalist who will find the profession London-based and closed to almost all but the independently wealthy. It’s printing replicated works of dead writers as we’ve done to Ian Fleming, Douglas Adams and Agatha Christie, without any help from technology. The key component of creativity is prompts and people, the interchange of ideas and a thriving ecosystem of composers, writers and artists with enough time and money to create experimental products. AI stands to replace the commercial work that many in the sectors use to make a living or start a career and that is a threat. But if it helps with ideas and frameworks as creators are increasingly time-poor and isolated, it is an opportunity. It’s a tool: the latest step in how technology has always changed how we make and share and spread ideas.

Enjoying The Critic online? It's even better in print

Try five issues of Britain’s newest magazine for £10

Critic magazine cover