If You Gave AI The Editing Rights…

A Thought-Piece on AI’s Creative Writing Editing Ability

Dorothy Paul
8 min readMay 3, 2024
https://www.istockphoto.com/photo/hand-with-pen-proofreading-gm508609021-45546846

Introduction

Being a creative writing editor is a multi-faceted profession. Not only do editors lead writers on what they need to revise in order for content to be ready for publication, but editors can also select content based off the parameters of the publishing firm. Essentially if a job of an editor can be summed up to one sentence it would be “I choose the perfect stories and make them more perfect.”

But of course, with a multi-faceted profession, the biggest enemy is time. I may not be a creative writing editor at a big old publishing firm, but I can only imagine that the workload of dealing with a long manuscript (sometimes multiple) to read in a day. But on top of reading also giving suggestions for revision can be very time-consuming atop of everything else.

As the title suggests, a potential solution to alleviate workload and shorten the time taken on tasks is using AI. But the question is: could AI be a good tool for editors to use in order to hasten productivity? How can AI push away some of the editing tasks that are menial and unimportant to the real job title? And what’s the best way to know how AI can help editors? By asking AI itself.

The Prompt: if you, as an AI, were to aid an editor at a publishing firm, what would you do to make their job easier?

Chat GPT gave me a list of ten tasks it could help an editor with. Here is a simplified version of the list:

  1. Grammar and Style Checking
  2. Language Enhancement
  3. Fact-Checking
  4. Plagiarism Detection
  5. Automated Formatting
  6. Content Analysis
  7. Trend Analysis
  8. Version Control
  9. Metadata Management
  10. Reader Feedback Analysis

For the most part, the AI listed tasks that would be the most surface level and less in-depth. For instance, to have AI work on the grammar editing while the editor themselves works on plot point editing, this would allow for the editor to essentially do two editing tasks at once. Traditionally, looking for multiple aspects to edit at once can be not only a pain but impossible.

Also, tasks like automated formatting to make the novel up to the publishing firm’s standard can cut down time that would otherwise be wasted by doing the task manually from scratch. Though scary, metadata management could also be important for AI to do as it defined that as “assist[ing] in organizing and managing metadata such as keywords, categories, and tags to optimize discoverability and SEO for published works.” What better tool to sway an algorithm into making work more discoverable than something similar to the technology that the algorithm is made of? This is also where trend analysis can come in handy.

So, these promises seem alright and even plausible on the surface level, what are some of the holes of these promises? Fact-checking is one promise that definitely has holes if you know about AI’s tendency to hallucinate. For instance, if there is harmful misinformation within a novel that is perhaps in a genre that’s more nonfiction (historical fiction, creative nonfiction, etc.), the AI may accidentally say it’s correct. That’s why in terms of fact-checking, it is best to read the information yourself, because in creative writing specifically, the misinformation may be intentional and in context, might be integral to the story’s characters and themes. Plagiarism detection could be another that could be a broken promise in practice. This is due to the fact that in other fields, specifically education, AI plagiarism detectors cannot detect certain forms of plagiarism, one of those being the use of AI in a written paper. But plagiarism and the forms of it are also nuanced. This is especially true when it comes to pieces that imitate others but cannot be proven to be actual plagiarism. The article Creative Plagiarism by Paula Marantz goes more in-depth, but the short of it is that spotting plagiarism becomes more complicated in creative writing as you never know when the author is perhaps writing an innovative adaptation of a piece, but not directly copying the piece itself.

How AI handles and understands creative writing and the unsaid practices is, of course, important to analyze when it comes to looking at editing for obvious reasons.

However, some of the promises do seem valid. AI already helps with grammar already in self-editing sites like Grammarly, so who's to say the practice can’t be refined to the point where it can accurately help with surface revision needs? Can the same be done with style as well? It could be entirely possible.

But let’s ask Chat GPT another question, shall we?

The Prompt: if you, as an AI, were to aid an editor at a publishing firm, what would you do to make their job harder?

Whereas before Chat GPT confidently gave a list of ten items it could help an editor with, Chat GPT only listed eight items on how it could make editors suffer:

  1. Overreliance on Automation
  2. Misrepresentation of Context
  3. Technical Errors
  4. Privacy Concerns
  5. Bias in Suggestions
  6. Resistance to Change — “[S]ome editors may be resistant to adopting AI technology or may feel threatened by the prospect of automation replacing their roles.”
  7. Complexity of Implementation — “Integrating AI systems into existing editorial workflows may require significant time, resources, and expertise[.]”
  8. Learning Curve — “Editors may need to invest time and effort in learning how to effectively utilize AI tools […] temporarily increasing the cognitive load on the editor.”

Lots of these issues can slow down an editor and not hasten their productivity. But what drew my attention was #6 where the AI stated that editors may be resistant to AI replacing their roles. I think the resistance to change would actually be an issue, but the way that AI frames it suggests that the sentiment that AI is a “job replacer” has become a popular enough one where Chat GPT addressed it here.

And based off its answers and what I have observed with AI myself, I highly doubt AI would replace editors. It even acknowledges how an overreliance on automation would hurt editors in the long run in this list.

The truth is: editors still do need to exist, and those skills are still important because AI cannot do everything. As it acknowledged itself, it can make errors and if there is no one to point out those errors, how would we know? Also, AI never said anything about helping editors choosing what content would be best based off the publishing firm’s standards. It only told how it can help with surface level tasks in editing and marketing that would make the job easier. The more in-depth tasks would still lie with the humans.

This is especially true when you look deeper into #2, misrepresentation of context. Novels are long and also one of the most popular mediums of writing. I have personally seen how AI misrepresents context which leads to some inaccurate suggestions that slow down the writing process; this was for a short story as well. I can only imagine how an AI would slow down the editing process for a novel with inaccurate style and grammar changes just because it misrepresented the context. Especially the former, considering AI doesn’t do a good job with creative style itself. And even if the suggestions for more difficult subjects such as plot and characters are good, we don’t necessarily know where they come from.

So as a tool for simpler tasks that wouldn’t muddle the whole process too much? Yes, AI could potentially help editors get through their simple but time-consuming tasks faster. Which would be a positive and give more room for the editor to work with the writer on more complex tasks and queries. Overall, giving more room to balance the writer’s vision with the publications’ expectations and marketing standards instead of worrying about more menial tasks.

However, there is an elephant in the room we haven’t talked about. An elephant in the room that has nothing to do with human editors for hire or editors at publishing firms.

So, What If AI Replaced Editors?

AI Editors (like this random one I found) are already something you can actively seek out and with the rise of self-publishing, there is a potential that people will avoid the hassle of paying for a human editor and instead use a free AI editor that can — allegedly — do the same job. And if you want to use a paid AI? At least you know the price going in. Depending on how a freelance editor charges, it can either be per word or per time spent on the piece. It varies.

So, why should creative writers pay for human editors when AI editors do exist?

An AI editor cannot “argue” with you about your creative choices, and it’ll be pleased with your every decision — unless it’s a surface level.

But, of course, an AI that is too agreeable and has problems we already discussed will be a weaker editor compared to a human editor who will actually give you valid constructive criticism. A human editor may not align perfectly with your visions in a way an affirming, programmed people pleaser would be, but a human editor does have one thing: it knows what human readers like and want based off experience the AI does not have. AI can try its hardest to replicate that expertise, but why use something that’s replicating when you can pay more for an actual human who is the real thing?

When I asked AI the same question: “why should creative writers pay for human editors when AI editors do exist?”

Chat GPT gave a list once again, but it ended with this sentiment:

Overall, while AI editors can be useful supplementary tools, many writers still value the depth of insight, personalized feedback, and human connection that human editors provide, making them willing to invest in professional editing services. (Chat GPT, 2024)

AI cannot replace human connection, feedback and insight. AI can enhance editor-to-writer communication by editing emails for tone or giving suggestions as to how to approach a tough conversation that may be life or death for the work relationship. But AI cannot replace the actual relationships themselves. Because of this, that’s why AI cannot give personalized feedback. Based on interaction, an editor can tell what compromises or changes can make a writer happy because they’ll have been working with each other for a long time thus get to know each other at a personal level. There is lots of complexity to relationships and that makes them deeper even if they are strictly professional. AI, for the most part, is sterile and just does what it thinks the person would like based off prompting, which can be just as sterile.

So, if you want a free AI editor? Go ahead. But you’ll be missing the human touch.

Conclusion

AI in the world of editing is a mixed bag of possibilities. For the most part, what really could happen to creative writing editors of all kinds due to AI is up in the air. Since AI, like Chat GPT, is evolving at faster rates than we can keep up, my predictions and its predictions can be moot by the time the next versions come out. But throughout these versions, one thing has stayed a constant: a pressing need to validate our importance as humans.

As illustrated through this article, humans have skills that AI can only dream of (if they dream at all). This is because we have real-life experience whereas AI only had fed experiences from the internet.

AI can be a great tool nonetheless that can help to decrease burnout and complete simpler tasks that don’t require lots of brain power but do limit the amount of work within a day. In fact, AI should expand our time to be more creative and innovative with our jobs and coworkers rather than take away the creativity and innovation aspects itself.

--

--