As technology advances, so do the tools available to creatives. First it was movable type, then mechanical pencils, and now we have generative AI. You’ve surely heard of it by now — it uses something called a large language model to create content based on prompts you give it. Generative AI tools can produce text (along with audio, video, pictures, and other things) and they can do it faster and cheaper than you, a human. 

That of course raises the fear that it could replace those of us who write for a living. And, yeah, that’s possible: if you write commodity copy, like paint-by-keyword SEO articles or super-safe marketing emails, generative AI will probably be able to do your job better and faster in the near future — if it doesn’t already.

But before you start studying for that real estate license, let’s recognize that, for most creatives — anyone whose job is to conjure original ideas and make something fresh — generative AI is just another tool you can use to streamline your process, or even to enhance your creativity. 

That goes for writers as much as anyone. We spoke to three who are already using generative AI writing tools in their work to get their insights on how to use AI to its fullest potential while protecting the human heart and soul in everything they make.

Plugging AI into your creative process

For novelist Leanne Leeds, generative AI was just a faster way of doing what she’d been doing already. “I kept a database of e-books I bought from writers that I liked. If I was writing and got stuck with phrasing or how to end a sentence or how to say something in a different way, I'd search through that database and flip through other authors' work to get ideas.” AI seemed like a more powerful version of her own database, so using it just made sense.

For anyone who writes, generative AI is a fantastic tool for getting out of ruts and finding the right word or phrase for the moment. Bestselling author Joanna Penn, who writes both thrillers and nonfiction — including “Artificial Intelligence, Blockchain, and Virtual Worlds: The Impact of Converging Technologies On Authors and the Publishing Industry” — uses it a lot for this purpose. “AI tools enable me to ideate and iterate faster, and to spark off into other ideas,” she says.

But it’s much more than that. “The most exciting part is that it opens up the creative process to everyone, not just a chosen select few,” says storytelling professor and consultant Jim Hull. Jim funneled his 20 years as an animator and director at Disney and Dreamworks into Subtxt, an AI writing assistant that pairs a predictive narrative framework with AI text generation, and he is “definitely bullish on AI supporting the creative process.”

He has seen many potential artists drop their creative pursuit because they can’t get past their impostor syndrome and just start making stuff. “They feel that if they weren't ‘born with it,’ or if they didn't go to art school that they're not allowed to be creative in any shape or form,” he says. “With these new generative AI tools, the neophyte artist can start with nothing — and quickly see something that, in essence, gives them that permission to go on.”

With our powers combined

But once you add AI into the mix, who’s to say what’s yours, and what’s the robot’s? If you can get AI to do your writing, art, or video production for you, why do you need to be there?

First of all, even the best AI isn’t good enough yet to fully replace the human hand. “I've yet to see something that has come back, whether through GPT-3, ChatGPT, DALL-E, or Midjourney that didn't need some kind of tweaking,” Jim says.

Joanna agrees. “There is no magic button I can press to output a perfectly formed finished product, nor would I want to, anyway.”

That’s the thing: however you use generative AI tools, the creative decisions — what to include and leave out, how to say what you want to say, when humor moments work, and so on — remain ultimately, entirely, up to you. 

“It’s like having a super car sitting outside your house,” Joanna says. “It has huge potential, and it can go in any direction and end up in any place eventually. But you have to decide where to go, and you have to drive it there. The tool is nothing without the human behind the wheel.”

In addition, most people don’t create in a vacuum — they have help. It’s just that their help is usually human. “Currently, my books are around 90% my own words and around 10% augmented in some way with artificial intelligence,” Leanne says. “That 10% that has been augmented isn't much more than what my editor would augment by cleaning up phrasings or clunky statements or rearranging the presentation for clarity.”

There’s also something deeper at play: most creators like creating. Ancient humans didn’t make cave paintings to sell at auction; they did it because there’s something inherent to humanity that drives us to create. While companies will inevitably start using AI to pump out blog posts, marketing materials, and maybe even a few creative efforts like songs and novels, humans are not going to stop creating — and by extension, they likely won’t stop consuming human-made content, either.

Shortcomings to keep in mind

So the good news is that AI isn’t perfect. That’s also the bad news, if you were expecting it to make creative work easy.

The first challenge for creators using AI technology is that many of these tools are really technical — though more and more products are coming out that harness the power of the underlying AI with a more intuitive user interface. For example, Joanna started by using OpenAI’s GPT-3 on its own, which she found too difficult. Then she found Sudowrite, which uses GPT-3 to analyze characters, tone, and plot and offer ideas for what should happen next in a story — “essentially a pre-prompted front end,” as Joanna puts it.

Leanne uses Sudowrite too, and finds that it’s less a silver bullet for her creative challenges and more like a slot machine. “You hit the button, pay a coin, and hope that something awesome hits that's perfect — but most of the time, that's not going to be the case. It's rare that I see a paragraph the AI generated and I'm instantly impressed and ready to incorporate the entire paragraph as is. It's happened, but it's rare,” she says.

Generative AI can also offer up creative suggestions you didn’t ask for and would never want. The AI behind Subtxt will sometimes try to rewrite fiction stories to make them happier, for example, despite the fact that some of history’s greatest literature ends in tragedy. This is something Jim is actively working to change so that the tool stays aligned with the artist's intentions.

Likewise, the AI might want to inject some action in a spot where the character is deliberately taking no action — to resolve a conflict, for instance. “This would be the part where human intervention is still needed, and where I see the generative process as additive rather than as a replacement,” Jim says.

And finally, there are the ethical concerns. AI is only as good as the data used to build it, and any data pulled from the world at large is bound to have biases. And the more AI-assisted content that’s out there, the worse those biases will become. AI can also get facts wrong or make things up entirely.

Then there’s the issue of plagiarism and copyright, which is a thorny subject. And one that writers can expect to be on both sides of — seeing their own work parroted by AI tools and learning that they lifted from another writer via AI tools. “We need to be a bit gentle with each other as a creative community while this is getting worked out legally,” Leanne says. “It's going to get murkier before the waters clear.”

There are a few fixes out there for these issues. Joanna worked with the Alliance of Independent Authors to create guidelines for using AI, which include things like being aware of AI’s biases; editing, curating, and taking responsibility for anything you publish; and labeling anything created with the help of AI.

“For example, I include a statement of AI usage in my books, and I add a badge on my AI-generated audiobooks,” Joanna says.

Should you be afraid of generative AI? 

Probably! Generative AI has the potential to usher in a generational shift in the way we do lots of things, including creating content. That’s scary, and it’s okay to fear it. But don’t let it stop you from experimenting — that alone could help you get past your fears.

"Most people who are scared have not tried the tools yet,” Joanna says. She advises starting with the tasks you don’t enjoy doing, like writing promotional copy or website metadata.

It’s also okay to be afraid for your livelihood. If you make a living doing something that a computer can suddenly do — well, we’ve all seen how that turns out. But if you’re good at what you do, it will be a long time before the AI can replace you. And figuring out how you can use it to make you even better is probably your best defense against it. Plus, fear is rarely an ingredient in any recipe for creativity.

“I don't believe what is ultimately fear-based thinking is good for anyone, least of all the artist,” Jim says.

And in the end, AI is coming — whether you’re ready for it or not. So creators might as well embrace it.

“It’s early days,” Joanna says, “but like the internet, or electricity, AI will soon be part of almost everything we do.”

The author used the AI tool Whiskey to turn her rambled thoughts into a cohesive introduction paragraph and Lex AI for suggestions when her writing flow stalled. All AI-generated content was then edited by a human.

‍