AI Creative Writing: Can AI Really Write Novels?

Artificial intelligence has made advances by leaps and bounds in terms of AI writing, leading to a rise in courses on how aspiring authors can use the technology in the writing process or even write entire novels. 

The real question is: Can AI handle creative writing? Can it really write a novel from start to finish?

Before you hand over the keys of your next bestseller to AI, it’s important to understand what AI creative writing can and can’t do.

Key Takeaways

• AI can rapidly generate text segments similar to what may appear in a story.
• However, AI can’t incorporate personal experience or emotions to add nuance that keeps readers turning pages.
• As LLM's capabilities expand to 128,000 tokens (approximately 250 pages), their capacity to generate longer sections of text for stories increases.
• Using AI to generate books poses copyright considerations as the U.S. Copyright Office advises that copyright only currently applies to a “product of human creativity.”
• Many genres pose specific challenges, such as world-building in science-fiction and the human experience in literary fiction.

AI Creative Writing: a Brief Overview

AI is trained on huge swaths of data, including websites, books, and more. Through its training, AI can learn to pull out patterns, styles, and concepts.

However, is AI able to create content that looks and reads like a novel? 

If you prompt an AI chatbot like ChatGPT to write a scene that might appear in a novel, it can typically generate a quick section of text that is similar to a passage from a book.

Yet, AI doesn’t understand emotions or tie in personal experiences the same way a human author can. 

For this reason, it can’t delve into the nuance and subtlety that create those layers of emotion that pull at our heartstrings or keep us turning pages. Sure, AI can write dialogue, but without extensive and specific prompting, it can’t create a deep, intertwined emotional story arc.

Can AI Write Short Stories or Full-Length Novels?

Does the length of creative writing impact AI capabilities? Let’s review using AI for short stories vs. AI for novels.

AI and short stories

You can prompt AI to write a short story, and tools like GPT 4o-mini will quickly generate a brief draft.

So, can AI create a short story? Yes. However, it won’t have the same impactful narrative inspired by human perspectives.

Similarly, it may even be prompted to generate individual chapters.

AI and novels

Can AI generate a full-length novel? It depends.

The actual length of a full-length novel can vary, and not all LLMs are equipped with the capacity to process that number of words or characters.

In their guide on context windows, IBM highlights that the LLM industry standard at the moment is 32,000 tokens. However, this is changing. Companies are looking to create and refine models that can handle 128,000 tokens, which is similar to the length of a novel that has 250 pages.

However, although the capacity for handling longer context windows may increase, creating a full-length novel adds layers of complexity that are best when human-written.

Great novels weave together:

• An intriguing plot.
• A unique setting.
• Memorable characters.
• Deep themes that keep readers engaged.

AI-generated novels can lose coherence, go off on a variety of trajectories, and leave gaping plot holes and loose ends. 

AI Creative Writing Challenges

There are a number of challenges that AI creative writing faces.

AI comes up short in creativity

When it comes to AI creative writing, it’s limited to its training.

Instead of trying to create new ideas, AI borrows a mix of themes or ideas it has encountered from its training. 

In terms of writing stories, its ‘plots’ become easily predictable — which doesn’t provide a great reader experience.

Additionally, if it generates any duplicate content, it could result in plagiarism or intellectual property issues.

Plot development

One of the biggest challenges that AI creative writing struggles with is creating and maintaining a coherent plot. 

Advanced models like ChatGPT can understand the basics of storytelling (when prompted) and grasp how to generate some key aspects of a story like a setting, conflict, and resolution. 

However, its novel-writing skills take a dive when weaving together the storyline layers that make an immersive novel like:

• Different subplots.
• Creating an understandable, logical progression from one chapter (or book) to the next.
• Expanding character progression to create compelling characters that react to plot twists and turns.

Character development

Human writers often spend weeks or sometimes months developing rich, varied characters with backstories, motivations, flaws and fears. 

Creating characters with this kind of depth makes a story resonate with readers. 

Readers can relate to human-crafted characters because they, too, know what it feels like to go through life similarly or encounter those types of obstacles. 

AI, on the other hand, can create interesting personas, but it struggles to maintain them.

Where AI is helpful, is in creating basic character sketches or brainstorming ideas. Then, a human writer can fill out the initial ideas to fully develop the character for the story.

AI and copyright considerations 

Further, according to current US policies, you can’t copyright AI-generated content. 

The United States Copyright Office, published a guide on content with AI material which notes, “copyright can protect only material that is the product of human creativity.”

So, generating a book wouldn’t provide the same protections under copyright as an original human-written work. Read more about who owns AI content.

Genre-Specific Challenges

What challenges do specific genres pose for AI creative writing? Let’s take a look at some of the genre-specific details that AI would have to master to truly generate a compelling storyline.

• Sci-fi and fantasy: Science fiction and fantasy novels rely on world-building (imagine a futuristic or magical society). This requires a deep internal logic to create settings that the reader can imagine.
  • AI struggles with hallucinations and can generate incorrect information. The same could potentially occur when generating a novel that relies on structured world-building.
• Literary fiction: Literary fiction often dives deep into stories that capture a very human perspective and experience and are rooted in psychology. What’s more, AI could incorporate clichés, making the copy feel robotic and formulaic.

Incorporating AI Into The Human Creative Writing Process

Although AI isn’t proficient at writing immersive novels on its own, there are areas where AI can contribute to the human creative writing process and save you a great deal of time. 

Consider using AI to brainstorm novel ideas

AI is great as a brainstorming partner that can suggest different character traits, various storyline arcs, or plot twists based on the prompts you give it. 

So, when you’re exploring different ideas that can make your novel come together, AI can offer helpful suggestions that help you beat writer’s block and open up new possibilities for you to explore.

Optimize your editorial process with AI

Some AI-based tools like a grammar checker can help review your text and make sure it’s grammatically correct.

This can help you improve the piece overall. However, it’s important to note that some features of AI editing tools that rewrite or rephrase text can impact AI detection scores.

Final Thoughts 

The bottom line is that AI is best as a support to original human writing when it comes to drafting a novel from start to finish.

There are areas where it can excel, including helping you overcome writer’s block, streamlining research, or even helping you diversify character’s personalities or voices.

AI can’t replace human creative writing, but it can help you take your writing ideas further. 

Looking for more ways to improve the content you publish? Try the Originality.ai Readability Checker to improve the clarity of your writing and make sure that it’s tailored to your ideal audience.