A silicon mind and human heart
Many years ago, I wanted to learn how to use Photoshop. I worked in advertising and sometimes needed specific images for my strategic presentations, rather than the standard ones I could find on the internet. I bought a book that explained Photoshop, but very soon found out that learning Photoshop through a book didn’t work for me. I then imagined a piece of software without a complex manual, that even I could use, which would make Photoshop obsolete – at least, for me. I was quite sure it would be invented at some point in the future. After all, most of the things that we imagine, are eventually built.
Nick Cave was not amused
With generative AI (artificial intelligence that produces text, images, videos, music, or other forms of content) today becoming an applied technology – rather than science fiction – I can now finally design my own images. That is, I can almost do so. It is still a somewhat clunky process, requiring lots of trial and error and often producing the wrong output. For the image above this article, I first tried Dall·E – the text-to-image app by OpenAI – but after several attempts realised it couldn’t translate my prompts in an acceptable image. I also tried ImageFX by Google. The result was more acceptable, but I couldn’t create a widescreen image. Then I tried Adobe Firefly, in which I uploaded an example – which was a big help. I still had to settle for a simplified version of what I imagined, but with so many emerging players in the field of generative AI – in what seems a Cambrian explosion – in a few years’ time translating my imagination into an image will probably be a piece of cake.
With so many emerging players in the field of generative AI – in what seems a Cambrian explosion – in a few years’ time, translating my imagination into an image will probably be a piece of cake
Not everyone is so enthusiastic. A year ago Nick Cave received the lyrics for a song from a New Zealand fan – called Mark – who asked ChatGPT (a generative AI that answers most of your questions) to write a song “in the style of Nick Cave.” Cave was not amused, calling it a “grotesque mockery of what it is to be human,” since “writing a good song is not mimicry, or replication, or pastiche, it is the opposite.” This is true. Creativity is about creating something that is original, not something that obediently colours inside the lines of what has already been created before. Which is what generative AI at this point does; it creates mashups of existing works. Understandably, as a true artist, Cave wants to create something entirely new. What’s more, being creative is also about going through a creative process that is sometimes frustrating and sometimes very enjoyable. Cave calls this “the authentic creative struggle that precedes the invention of a unique lyric of actual value”. Prompting a machine with ‘in the style of’ is too easy, in Cave’s view.
Combining ideas and styles
I understand where Nick Cave is coming from. But we should be careful to look down upon ‘in the style of’ or a mashup of existing works. After all, when the human brain is creative it also combines existing ideas and styles into new creations. And when Nick Cave is writing songs he uses – both consciously and subconsciously – sounds, words, and lyrics that were already used by others. He is also inspired by other musicians, writers, scenes, and random ideas. Since the beginning of music, people are making music that resembles someone else’s style.
Clearly, the ChatGPT-lyrics artistically don’t come close to the poetic and sensitive writings of Nick Cave. But since AI’s mechanisms of deep learning and combining ideas are similar to that of the human brain, I am afraid it won’t take long before the distinction between human and artificial creativity is much harder to make. Especially when AI reaches the level of artificial general intelligence, but this will probably still take a few decades.
I already came across some pretty qualitative artificial writing in the magazine Poets & Writers
I already came across some pretty qualitative artificial writing in the magazine Poets & Writers. In this magazine Eileen Pollack, professor at the University of Michigan who has been teaching creative writing for 20 years, wrote an essay titled The Antithesis of Inspiration: Why ChatGPT Will Never Write a Literary Masterpiece. The reason for her article is that people around her were constantly asking her whether “our best creative writers will be replaced by their artificially intelligent counterparts”? To answer that question, she spent a few weeks playing with ChatGPT. And she “was stunned by the quality of its output.”
A delicate web of emotions
In here essay Pollack entertainingly describes how she prompted ChatGPT in different ways and what its subsequent output was. For example, she asks ChatGPT to compose a paragraph in which the narrator describes her parents’ obsession with maintaining neatness and order in the 1950s ranch house in which she grew up. “To my amazement,” she writes, “the program reeled off a paragraph that would warrant high praise if one of my undergraduates turned it in. The grammar is perfect, the syntax varied; the language, if not sparkling with originality (“picture-perfect dwelling”), isn’t riddled with clichés.” Nevertheless, as the title of the article announces, Pollack is convinced that ChatGPT will never write a work of literary genius.
With things like the internet, mobile apps, augmented reality, and AI becoming more and more integrated in our lives – and at some point, even our bodies – we are surely becoming digital creatures.
The reason, according to Pollack, is that AI-program cannot simulate “a delicate web of emotions” and “will never be inspired by its own embodied experience in the unpredictable physical world to experiment with new literary forms or content or to seek answers to previously unexplored questions.” The distinction between the physical and digital world, however, is becoming less and less clear. For example, when we visit a city and use Google Maps for orientation, we are already putting a digital layer on top of the physical world. And when we listen to Discover Weekly on Spotify, we are feeding our taste with the music an algorithm serves us. In fact, we might be listening to AI-generated music. When we pick up our mobile phone a hundred times a day, we are also constantly being digitally inspired. With inventions such as the internet, mobile apps, virtual reality, and AI becoming more and more integrated in our lives – and at some point, even our bodies – we are surely becoming digital creatures. If we aren’t already.
AI is merely a tool
Besides the inevitable digitalization of our world, we also shouldn’t forget that for the creative mind, AI is merely a tool. And tools are used to enhance our human skills. Pollack could be right in predicting generative AI to never write a masterpiece. But when AI is used as a tool, it can be used as our researcher, creative sparring partner and editor. Which means that those who wouldn’t receive too much praise in Pollack’s creative writing course, but have the talent and ambition to tell great stories, could publicize a book with the help of AI. Mind you, a book that will be valued by the gatekeepers of the literary domain. And while AI might not even be mentioned on the cover as the creator, the manuscript will be the fruit of a symbiotic relationship between the author and his or her AI-powered ‘co-writer’. So, to be truly creative (i.e. original) generative AI will require human inspiration, in the form of original plot summaries, visions, deep emotions, strange characters, mind-blowing scenes, subtle jokes, random events, and – let’s not forget – ‘in the style of’ prompts. And by constantly adjusting the output – and maybe even hiring a professional human editor to do so – the result might very well be a (collaborative) masterpiece.
To ask whether generative AI will be able to create a work of literary genius, is a little like asking whether the screenplay writer could create a blockbuster
The described collaborative relationship between human and machine could be compared with a director creating a film with a team of specialists. To do so, he or she works with a screenplay writer, cameraman, actors, set builder, digital designer, etc. Without this team, the end result would be pretty meagre. To ask whether generative AI will be able to create a work of literary genius, is a little like asking whether the screenplay writer could create a blockbuster. Both AI and the writer are at best co-creators. They might enrich the creative process with their own creativity, but the ‘director’ (or puppeteer, if you like) is primarily responsible for the result. Therefore, ChatGPT should be seen as an assistant-writer, which, for that matter, will make the creative process for many less of a struggle.
More mediocrity
When AI is being used as a collaborative tool, maybe the more interesting question is; how will this technology influence human creativity? As said, this is hard to predict, though some of the contours are vaguely becoming apparent. Pollack expects that “lazy students, hack writers, and cost-cutting bosses” will use AI to create “mediocre essays, novels, newspaper articles, and television shows.” Probably. There’s more, two few weeks ago 200 artists signed a petition, asking tech firms not develop AI tools that replace human creatives. Generative AI might indeed kill some (creative) jobs. Will there also be an upside to these job-killing algorithms spitting out mediocrity? As with all industrial revolutions; probably yes. The reason is that there is also a very plausibel mechanism that speaks against an overdose of mediocrity. If we can delegate some of the creative work to a machine, and if this machine can compensate for some of the creative skills many of us lack, the average level of creativity might just as well increase. In which case we might even create more masterpieces.
What AI allows for is that those who have a creative mind and strong motivation to write or compose a masterpiece, can more easily compete with the Philip Roths and Beyoncés of this world
More masterpieces doesn’t mean that Mark from New Zealand will suddenly become an acclaimed singer-songwriter. Or that I will become a successful art director. You will always need the talent and intrinsic motivation to become one of those things. But what AI allows for is that those who have a creative mind and strong motivation to write or compose a masterpiece, can more easily compete with the Philip Roths and Beyoncés of this world. Though I realise that ‘more easily’ sounds somewhat misleading. Creating a masterpiece will never be easy. In fact, becoming a successful artist requires more qualities than just writing or composing. To name a few: self-confidence, persistence, being social (within a network), and being able to promote one’s work. So, for those who crave the creative struggle, there’s still enough to worry about.
With AI on board, the aspiring musician, who might not be as musical or articulate as Nick Cave, can collaborate with different kinds of AI-fed software to raise his or her level of creativity. In fact, even Nick Cave could. You can compare it with how the personal computer – and all the inventions it enabled – allowed for new forms of creativity. And, while perhaps making certain jobs redundant, it also created many new jobs. In the light of this comparison, AI suddenly sounds a little less revolutionary. It is yet another tool that enhances our creativity – rather than kill it.
Cyrano
Eileen Pollack ends her essay with an anecdote about herself dating a mathematician. After some flirty texts, he sends her “a long, beautifully written story” about someone who yearned for “someone who could share in his love for numbers and the secrets they held.” Pollack was impressed by the scientist’s ability “to compose such a witty, charming story”. Until she “realized he must have used ChatGPT as his Cyrano”. She felt cheated and explains that she has “no desire to connect with the silicon mind and digital heart of an AI program”. This surprised me. You might argue that it was a silicon mind who wrote the story, but it was the mathematician’s heart who ‘directed’ it. While some artists might never get used to the idea, we can expect many generations to come to evolve into more creative – albeit more digital – creatures.