AI vs. the Human Copy Editor

 This rise in the widespread availability or popularity of AI in editing spells trouble for the copyediting industry, or does it? Many professional copyeditors have been left anxious with the rise of LLM-based AI software. 

These large language models have been trained on vast amounts of human data and can both write and edit content in minutes or seconds, where a professional writer or editor would take hours or days. 

Such tools are cheap and in many cases free, or at least they have a freemium tier where users don’t have to pay to use. Of course, this presents a perfect opportunity to cut costs for businesses and anyone who requires the services of an editor. 

According to the Editorial Freelancers Association (EFA) — the largest and oldest professional organization in the US — copyediting prices and proofreading rates can range between $0.03 to $0.06 per word or $40 to $60 an hour. This cost goes down to zero and even close to zero when one takes into account the relatively low price of the premium version of AI models.

While LLM-based AI models have proven to be highly popular among the general public, the common complaint is that they are not good enough for professional-level work in both the worlds of academic or scientific writing and professional business copy. 

In the academic world, we saw a dramatic example of this with the fate of Professor Paul Yip, who had to resign from his position at the Faculty of Social Sciences at University of Hong Kong after it was discovered that he co-authored a paper that included several fabricated AI citations. Similar events have taken place in the corporate world.

For example, the consulting firm Deloitte promised to issue a partial refund to the Australian government after it was revealed that a $ 44,000 report was riddled with artificial intelligence-generated errors, such as made-up references. So what is it about AI and how it works that makes it so liable to making these types of errors? 

What’s wrong with AI?

One of the major problems facing AI is what is called hallucinations. This describes a process in which AI chatbots just make up information when they don’t know the answer. And what makes this problem even worse is the confidence with which they provide these answers.

AI systems were designed to give the most likely correct answer based on the information that they have available. In many cases, this means that they are simply unable to admit that they don’t know. This type of uncertainty is potentially disastrous for professional settings. 

For example, in the medical field, hallucinations about things like medical prescriptions cannot be “almost” right. The answer has to be definite. The same can be said of services such as accounting or psychology. This means that, ideally, a human copyeditor who knows how to fact-check information would be a necessary part of the loop when it comes to AI workflows. 

Lastly, AI chatbots are — well — robots. They simply do not know how to communicate with a target group of readers in the right tone. This is crucial in content marketing. The point of content marketing is to sell, and to sell, you need to build trust, which cannot be done through the hollow tone of a robot. In fact, there is even a whole new subindustry of copy editors being hired to edit content that has been written exclusively with AI. 

---

Cite this EminentEdit article

Antoine, M. (2026, February, 07). AI vs. the Human Copyeditor. EminentEdit. https://www.eminentediting.com/post/ai-vs-the-human-copy-editor