SPECIAL FEATURE The case for creative AI (Part 1)

 

Will Artificial Intelligence (AI) take our jobs or enhance the way we work? Christian Darkin reports.

If you haven’t tried out these creative AIs, you should probably stop reading right now, and visit:

Chatgpt - https://chat.openai.com/chat if you’re an author

Or Stable Diffusion - https://huggingface.co/spaces/stabilityai/stable-diffusion if you’re an illustrator.

You can come back when you’ve picked your jaw up off the floor…

 

It’s now possible to type a few words into an artificial intelligence like Dall-E, and have it produce a painting, a sketch or a photograph of literally anything you can imagine. You can also visit another AI; ChatGPT, and ask it to write a story for you.


How should we, as creatives, respond to the sudden emergence of revolutionary creative AI which, on the face of it, appear to have the potential to threaten the livelihoods of all of us and make us question what it is to be human?


Clearly, fear is one justifiable instant reaction. The world is about to change forever and in ways we cannot predict. However, we should be excited too. We are the creatives, and it is our job to shape the change.

AI will not take your job. Someone who knows how to use AI will take your job

Think of it this way: When photography first appeared, a lot of artists were very upset. What is the point of painting, they said, when you can just point a camera at a landscape or person and get a perfect reproduction? Art, many concluded, was dead.



They were about as wrong as it is possible to be.



What actually happened was that photography flowered into art-forms all its own. Artists took cameras and made amazing, incredible images. Photography was used to capture things that painters could never have captured. Cinema appeared, and with it, animation - a whole new outlet for visual artists.



And once painters accepted the idea that art was about more than just trying to make accurate representations of nature, there was a creative explosion. Surrealism, impressionism, cubism, pop-art, abstract painting and collage all flowered (arguably) in response to photography.



150 years on, there are more artists working in traditional media than ever before. Crucially, virtually every one of them uses photography in their process - whether that is for reference images, for inspiration, or for the final reproduction of their work on screen, or on paper.



In the same way, no artist or writer today can afford to ignore AI.

The Plagiarism Debate


The first response of many artists has been defensive - The advent of creative AI has caused a storm in the creative community. Millions of pages of text and visuals have been scraped from the Internet, and used to train powerful AI models to understand what humans mean by writing and pictures. This data includes the creative work of writers and artists, and it’s been harvested without permission. This has made a lot of people very angry.



While understandable, I would argue that this opposition is misplaced and that in any case, arguing about it will not stop the AI revolution. I suspect this for two reasons: one philosophical, and one pragmatic.



The philosophical reason is that when an AI looks at a painting, or reads a book, it is not memorising it. It is not making a copy in its mind - the “mind” of an AI is far smaller than the data it’s trained with, and it doesn’t have a “memory” as such. A neural network is a digital version of the neurons in a human brain. It looks at art in exactly the same way that an artist or a writer looks when learning their craft. It is seeking to understand composition, brushwork, juxtaposition of subject and colour, characterisation, plot, voice and atmosphere. It learns as an art student or a writer learns. It is not stealing the work it views any more than I am stealing the Mona Lisa by viewing a photo of it on a webpage.



Sometimes, when it produces art, that work can be startlingly original, and sometimes it can be extremely derivative (this is largely dependent on the skill of the person prompting it). Sometimes, it can be prompted to produce works which are very similar to artworks it has seen many times before. This is just as true of most human artists and writers as it is with AI, and copyright rules are in place already to deal with this.

Crucially, the copyright police don’t turn up when a young artist or author arrives at a library or an art gallery hoping to soak in the culture. They don’t turn up to police that artist or writer’s imagination, or their notebooks, or even the art they create. Copyright rules are broken only at the point of sale. Inspiration is not the same as plagiarism (as anyone out there writing a story about a young wizard, or a vampire, or a detective, will know), but the two are on a continuum.



It’s only when you try to sell your work that you can be accused of copyright breach. I for one, don’t think art is helped by stopping it from being looked at. If we’re going to have AIs making important decisions that affect us humans (and we definitely are), I want that AI to read our literature, to look at our art, and to listen to our music. It’s the only way AI is going to understand what we are.



The pragmatic reason is that money is pouring into AI from every direction, and it’s easy to see why. AI is getting exponentially more powerful, and exponentially cheaper, and any company country or individual that ignores it is going to be left behind. This revolution is going ahead whether we like it or not.



We creatives are a tiny part of the workforce, and AI is coming for everyone. It doesn’t matter if you’re a doctor, an accountant, a civil servant or a teacher, a truck-driver or a farmer. The whole human workforce will see our work either change beyond recognition or vanish entirely in the next decade or two. We can either stand on the shore trying to order back the waves, or we can grab a surfboard and see where it takes us.



Creative AI is also un-policeable. There is no practical way to tell which art and literature is created with the help of AI or what data those AIs are trained on. In addition, most of the AI software is already freely available. Anyone can train a creative AI on any material they like, and nobody will ever know.

Is AI a creative tool or an artist?


The answer to this question isn’t simple. It’s likely we need to consider it as somewhere between the two. When AI produces words or pictures, it’s using a process which is very similar to that which human artists use. However, it only ever creates (currently) when it is prompted to do so, and it’s affected very heavily by how the person prompting it works with it.



We need to get away from the idea that co-creating with AI is simply typing in a prompt, waiting for it to spit out a story or a picture, and printing the result. That’s what people playing with the software will do but it’s not what artists do. Artists have a process, and that process can include a whole range of imaginative disciplines, research and reference work, experimentation, editing and refinement, along with technological tools, media, and technical skills.



AI can fit into an artists’ process in a thousand different ways. When you search google for a reference image, you’re using AI. When you notice an interesting story on your social media feed, you’re using AI. When you accept the corrections of grammarly, or a spellchecker, you’re using AI. Creative artificial intelligence is here to stay. The only question for writers and illustrators is how we choose to interact with it.



Saying work is “created by AI” is no more descriptive of the work than saying it’s “created by a pencil”. We need to have a more nuanced response about how and where AI fits into each creative process. Was the work we’re looking at the result of a single prompt? Or did an artist spend weeks refining their images, editing and combining a thousand variations, sketching, and re-working, adding hand-drawn details to create something truly original?



Around World Book Day this year, I visited a lot of schools. In every class, someone asked the question “Where do you get your ideas from?” I usually reel off a list of sources. Next year I may add AI to the mix.

Stay tuned for Part 2: Next week we’ll have some practical ways in which the new technology can be brought into the practice of artists and authors. This month, Christian tackles some of the opportunities and myths around the new tech.

*Header: Christian's test image using AI. All images: courtesy of Christian Darkin

*

Christian Darkin is an author, illustrator and animator who has been following the rise in AI since writing about it two decades ago for The Times and The Guardian. In a two part series, he presents a (possibly controversial) positive view of the present and future of this technology. 

*

Stephanie Cotela is the Network News & Events Editor at Words & Pictures magazine.