I’ve written before about why I don’t think AI is going to replace freelance fiction editors, but it’s time to revisit the topic and dive a bit deeper.
My main argument previously was that AI does not have emotional and creative judgement, and those things are incredibly important for fiction editing especially.
Those things are still true. But AI tools are constantly improving, and editors are increasingly being told they need to use them or risk falling behind. So I feel it’s worth revisiting the topic with a more practical lens.
In this article, I want to further explore three key questions:
Can AI tools help professional freelance fiction editors?
Should professional freelance fiction editors use AI tools?
Will AI replace human editors?
Let’s discuss.
Can AI tools help professional freelance fiction editors?
There are many AI tools available these days, and I’ve been experimenting with a handful of them over the past few years. I’ve also had my ear to the ground, keeping up with what other people in the industry are reporting.
And you know what? I honestly believe AI editing tools are completely inadequate.
Sentence-level editing: still unreliable
- They get things wrong more than they get things right. Numerous editors have put AI to the test. Here’s Adrienne Montgomerie’s series doing just that. Explore it for yourself. Grammarly is particularly abysmal, introducing way more errors than it corrects.
- They flatten voice and eradicate creative risk-taking. These tools were never really built with fiction and creativity in mind. They apply blanket rules, with no understanding of context or purpose. AI has its own ‘voice’ – which is why people can spot AI writing. Why would any creative writer want to sound like a machine?
- They don’t understand the rules they attempt to apply – let alone how to break the rules for effect. I’ve asked large language models (LLMs) such as ChatGPT and Claude to explain certain grammar rules. At first glance, they produce something that looks plausible. Dive deeper, and their explanations fall apart. Ask them to apply the rules to a piece of text and come up with examples of the rules in action, and they often produce something incorrect. AI may be more sophisticated than autocomplete, but it still has no genuine comprehension. It has no sense of logic and no understanding of nuance. It doesn’t know stuff.
No matter how many errors AI editing tools catch, they are so unreliable that it often takes more time sifting through their suggestions than it would to simply trust your brain as a trained editor. And that’s not to say anything about the judgement creative nuances require …
Developmental editing: a poor fit
Some people suggest that editors can use AI to do these developmental editing tasks:
- Conduct the first read-through, providing chapter summaries.
- Compile character lists.
- Analyse how well a manuscript has used a particular story structure.
- Make suggestions on strengths and weaknesses of the story, and potential rewrites.
These suggested usages make me very uncomfortable.
First, outsourcing something like the first read-through seems totally pointless. The human editor needs to read the manuscript in order to understand and engage with it. A summary just doesn’t cut it.
AI is infamous for not being hugely accurate and for ‘hallucinating’ (i.e. inventing) things. And so if you have to check everything carefully (which you do), the time-saving argument starts to fall apart.
Also, I’m against the creation of deliverables for the sake of deliverables, so any character list I create as developmental editor needs to be customised for a particular reason, and I’d need to work through the manuscript to understand what details I’d want to include here.
As for analysing structure, strengths and weaknesses and suggesting rewrites … Why would someone pay an editor for this when they could just go straight to the AI?
Remember: AI gets things wrong, and it flattens and homogenises. Even using it for a jumping off point would be potentially detrimental, because you’d be deferring your emotional human response to a creative work to something far more sanitised and mechanical.
If I were an author and I received a developmental edit that was founded on AI output – even if the editor reassured me that they had assessed it – I would feel deeply disappointed and my trust would be gone.
What about using AI for content and admin?
As you can see, I don’t think AI tools are very helpful to editors when it comes to the actual task of editing. But what about other usages – writing blog posts, social media posts and emails?
I can see the appeal: it would save a lot of time.
But the results can be damaging to your brand and trustworthiness.
People are becoming more and more attuned to AI-written content. As soon as I notice something I’m reading has been written by AI, I get angry. I feel tricked – and undervalued. There are also the inaccuracies and homogenised voice to contend with, both of which undermine credibility.
That said, there are a few areas where I’ve had some success using AI in my business:
- Macros. Getting AI to write scripts for specific, fiddly, mechanical tasks is a low-risk, high-impact use I can easily check and put straight to use.
- Blog audit. Feeding data downloaded from Google Search Console into an AI tool has given me useful, actionable insight with minimal effort.
- Search. I’ve started using Claude or ChatGPT in place of a search engine when I want something very specific – the results are often more targeted than Google, though false positives do appear, so everything still needs verifying.
I’m sure there are other genuinely useful applications. But I strongly discourage using AI for editing or client communication – unless the ideas are entirely your own and you’re prepared to edit the output extensively so that it sounds like you.
So to answer the question ‘Can AI tools help professional freelance fiction editors?’ – my answer is: not really.
Should professional freelance fiction editors use AI tools?
The ethical dimension
Some people have a hard stance against AI and won’t touch it with a bargepole due to ethical reasons: the way most AI models are built by flouting copyright laws, the impact on the environment, the fact it perpetuates damaging biases and so on.
I respect that.
Those things bother me too. But I am still personally open to using AI tools because to live in this world – in a capitalist society – means making ethical concessions every day, from the food I choose to eat to the companies I buy from.
There’s a whole (very large) discussion to be had about this, but here is not the time or the place.
Keeping up with technology
Besides the ethical dimension, there’s the argument that editors should be using AI tools in order to keep up with modern technologies and not become obsolete.
I believe companies who sell AI tools – and people who sell AI training – play up to this more than is necessary.
As I’ve just discussed, I don’t think AI tools are actually helpful to editors, at least not yet. It’s possible they will be in the future, but the rate of progress is slower than AI companies (and those with investments in them) would have you believe.
What I do think editors should be doing is keeping a finger on the pulse – reading up on how things are developing and experimenting with these technologies in a way that doesn’t break trust with their clients.
Mignon Fogarty (aka Grammar Girl) has a detailed and fascinating newsletter that I recommend: aisidequest.com
Will AI replace human editors?
As someone who provides editorial training, I don’t want to be putting time and effort into a business that’s becoming obsolete. The fear that AI will take my job is one I share with many others, including many freelance editors.
I don’t have a crystal ball. Some people think it’s inevitable that AI will take huge swaths of jobs – we could experience something much like the industrial revolution, with a large portion of people out of work while the machine-owners profit. Others think AI is a bubble that will burst as it fails to live up to its promised potential.
I suspect things will land somewhere in between.
What might actually change – and what probably won’t
I predict that AI will replace some editors. In fact, that’s already happening.
Clients where price trumps quality have already turned to AI for editing (and many other tasks). And I’m happy that AI tools can assist those who would have no budget for professional services anyway. Hopefully they find some value in it.
But editing with existing AI tools still needs an actual human editor to judge the output and shape it into something useful and meaningful – so much so that most of these tools don’t fulfil the time-saving promises they make.
Where I do think AI will improve is in mechanical editing – spotting grammatical errors, punctuation slips, consistency issues. The more rules-based and repetitive the task, the more tractable it becomes for a machine.
But the nuanced, emotional work of understanding a story or a line of resonant prose, and knowing what a particular author needs to hear and how to say it? I don’t think AI will ever get there – because of the very fact it is not human.
The question of client preference and disclosure
It’s worth addressing what authors and publishers actually want when it comes to AI.
Authors who are enthusiastic about it tend to use it themselves – and often bypass professional editors altogether.
Authors who seek out human editors are, more often than not, doing so precisely because they want that human connection. Many find AI involvement in their manuscript deeply uncomfortable, for reasons that go beyond practicality.
Fiction is about connection – one human’s imagination reaching another’s. For a lot of writers, readers and editors – me included – that feels sacred, and worth preserving at every stage of the publishing process.
Publishers are less consistent. Some are firmly against AI tools; others are more open or simply haven’t formed a clear policy yet. If you work with publishers, it’s always worth raising the conversation explicitly rather than assuming.
What I think is non-negotiable, regardless of where you or your clients land, is disclosure. If you use any AI tools in your editing work, be upfront about it. Trust, in this industry, is everything.
If you’re a new or aspiring editor…
If you’re just starting out, all of this can feel a bit daunting and confusing.
What’s worth holding onto is that the core skills of editing haven’t changed. Being able to read closely, think critically, understand story and respond with care – those are still the things that matter.
AI doesn’t remove the need for those skills. If anything, it highlights their importance.
Closing thoughts
AI will keep evolving. What’s less certain is whether greater capability will ever translate into the kind of emotional intelligence that fiction editing actually requires.
For freelance fiction editors, I recommend you keep thinking about where AI might fit (and not fit) in your processes – and to keep deepening the skills that no algorithm can replicate: emotional response, genuine comprehension and the sensitivity to serve the work and the author.
And then there’s the question no one can quite answer yet: even if AI does improve, will the majority of the people who work in this creative industry want it to be involved? Will the authors, the editors, the readers?
Fiction exists because humans want to share their inner lives with other humans. That impulse isn’t going anywhere. And I think that matters more than any technological capability.
FURTHER READING
I found the following blog posts particularly interesting and informative:
Why generative AI isn’t a useful tool (right now) for expert book editors (The Narrative Craft)
Why AI won’t replace human editors – and AI agrees (The Better Edit)
Copyediting and AI: a manifesto (The Better Edit)
The Hidden Costs of AI Copyediting Tools: An Editor’s Review (Jane Friedman)
Interested in training as a freelance fiction editor? Take a look at my online editing courses.
