Where’s the voice?
You know that feeling of opening a new book, reading a few pages, then finding you’ve been drawn into the story by the lilt of the author’s tone, their turn of phrase, the skill in the mood they’re creating and the intricate world they’re building? That, my dear, is an authentic voice drawing you in. There’s none of that in GenAI writing. It’s stilted, inauthentic and often overwritten.
There! I’ve said it.
So now, here’s the why.
AI writing is generic
The crunch
LLM output is derived from crunching text from thousands of scraped works, then spitting out the most frequently occurring sentences and phrases, resulting in an obvious generic pattern.
The problem with this pattern – apart from containing invisible embedded watermarks, now required by EU law – is that the prose is mind numbing for readers. This means, readers unable to do the one thing they crave: immerse themselves in an authentic human, storytelling voice – whether that’s warm, wishful, sharp, poetic or clinical.
The whole point of reading stories is immersion in someone else’s world.
Why AI writing is inauthentic
While GenAI output is useful in the corporate world for short-form texts such an emails, summaries or reports where we would expect it to be generic, in long-from works such as novels and memoirs we require ebb and flow, high tension and low, simple and deep expression of the human condition by a unique voice. Instead, AI provides the most common, cliched, static, jarring and over-written prose – because of the crunch.
Sometimes, this prose may look clever, clean and impressive – since it’s been stolen from some talented authors (and some not so talented) – but once parsed, it’s often wordy and repetitive. Weird and nonsensical metaphors are also a frequent feature.
This type of prose is exhausting for readers. The occasional complex and strange sentence requires reading and re-reading to understand what the heck is being said, only to turn out to be a garden-path sentence that goes nowhere, while the remainder of the text is like eating the same tasteless meal for breakfast, lunch and dinner, for weeks on end, without any variation in texture or flavour.
Texture and flavour come from an author’s unique voice and perspective – the way they express themselves through the granular use of words and their arrangement (syntax). This includes vocabulary, phrasing, tone/mood, rhythm, pacing and punctuation. All of which originate from the author’s unique perspective of the world they have created.
Like learning to play a musical instrument, an author develops their unique voice by practising. In the literary sense, this means reading, reading, reading and writing, writing, writing, then rewriting. There are no shortcuts to developing a unique expression. It needs to be organically built over time.
A consistent author’s voice earns trust from the reader. It shows that the author cares about their creativity, and isn’t just generating pap to take a shortcut to publishing.
And if a writer thinks that their big idea, theme or plot points are more important than the words on the page, they’re kidding themselves. Anyone can come up with an idea. Stories are not unique. They’ve all been told before. It’s how the story is told that makes it unique. Getting the story onto the page using your own voice is what makes it unique. It’s doing the work.
After all, if you can’t be bothered writing your own story, why should a reader give you their money or time to read it? And why should an editor waste their energy trying to make a manuscript out of mud?
Why texture and flavour are so important
Texture and flavour come from an author’s unique voice and perspective – the way they express themselves through the granular use of words and their arrangement (syntax). This includes vocabulary, phrasing, tone/mood, rhythm, pacing and punctuation. All of which originate from the author’s unique perspective of the world they have created.
Like learning to play a musical instrument, an author develops their unique voice by practice. In the literary sense, this means reading, reading, reading and writing, writing, writing, then rewriting. There are no shortcuts to developing a well-honed voice. It needs to be organically built over time.
Once that voice is established, except for the occasional superbly expressed line that may take a reader’s breath away, the words themselves tend to disappear for the reader – they simply unconsciously absorb the voice, allowing them to relax into the story and focus on scene and characters.
This is the point where a consistent, authentic author’s voice earns trust from the reader. It shows that the author cares about their creativity, and isn’t just generating pap to take a shortcut to publishing.
And if a writer thinks that their big idea, theme or plot points are more important than the words on the page, they’re kidding themselves. Anyone can come up with an idea. Stories are not unique. They’ve all been told before. It’s how the story is told that makes it unique. Getting the story onto the page using your own voice is what makes it unique. It’s doing the work.
After all, if you can’t be bothered writing your own story, why should a reader give you their money or time to read it? And why should an editor waste their energy trying to make a manuscript out of mud?
The idea is secondary
If a writer thinks that their big idea, theme or plot points are more important than the words on the page, they’re kidding themselves. Anyone can come up with an idea. Stories are not unique. They’ve all been told before. It’s how the story is told that makes it unique. Doing the work, getting the words down onto the page, using your own voice, is what makes it special, memorable.
After all, if you’d rather take the easy route with AI because you can’t be bothered writing your own story, why should a reader give you their money or time to read it? And why should an editor waste their energy trying to make a manuscript out of mud?
Read more on this subject
Part 1: Why GenAI causes publishing copyright issues: GenAI Risks Part 1
This blog explains exactly how GenAI works and why publishers are turning away manuscripts written with it.
Part 3: What GenAI tells look like: GenAI Risks Part 3 (coming soon)
This blog provides a list of examples that flag typical AI writing.