A copyeditor holds the final red pen before the text is published. It is mechanical work. This phase of editing is wherein the copyeditor searches for technical errors, consistency, and verifies facts. Copyeditors are stern to uphold the rules of grammar and must keep in mind the work’s audience and its intentions: the author, the reader, and the publisher. According to The Copyeditor’s Handbook, “all these parties share one basic desire: an error-free publication…the copyeditor acts as the author’s second pair of eyes, pointing out—and usually correcting— mechanical errors and inconsistencies; errors or infelicities of grammar, usage, and syntax; and errors or inconsistencies in content,” (Einsohn and Schwartz 3). This means that the copyeditor is only as good as their eyes are keen, or the text is looked at with a second opinion. The Copyeditor’s Handbook acknowledges that humans are faulty: Every copyeditor misses errors here and there; according to one study of human error rates, 95 percent accuracy is the best a human can do (Einsohn and Schwartz 5).

​

The copyeditor’s task is an essential step in the publishing process and acts as the periphery to the manuscript or draft copy becoming a published book or article. In decades past, copyeditors or an editor would need to correct text by hand, accompanied by their publisher’s style guide manual. The industry has long surmounted the days of handwritten manuscripts and much of the work is efficiently done on computers. This is known as on-screen editing or electronic manuscript editing (EMS). With the innovation of this technology, the software of Microsoft Office, Google Workspace, or Adobe Creative Suite are where writers and publishers produce work. These platforms’ applications each provide an assistive editing tool, a modern convenience that publishers may depend on to cut time in half. This may not replace the skills of a professional editor, but it does draw attention to minor mistakes of the text file owner, underlining spelling errors in red and underlining grammar suggestions in blue.

​

Copyeditors are expected to line edit, mechanical edit, language edit, markup, query content or structural edit changes, double check author permissions to referenced text, and consistency.

​

Artificial Intelligence (AI) and machine learning (ML) algorithms can mimic and create text with any given input. Companies are currently outsourcing their blog posts to services like Unbabel—which uses AI systems trained on text samples from actual human copywriters (Nizinsky). Nizinsky reveals at the end of his article that it was generated with the use of Copy.ai, commonly used for marketing copy. It is predicted that AI will be the preferred mode for marketing in the future, as it already is used by companies to generate social media posts and articles, as proven by Nizinsky. The reason why marketing as a practice is interested in AI is because it will revolutionize engagement, customer service and automation—three key objectives of marketing success (Melotti). Since the work of publishers is done digitally, it would not be difficult to adopt the practice of correcting text with AI. Once fed the publisher’s style guide, it corrects the style and tone to match, in seconds, eliminating the need for the human copyeditor. It is shown that it is difficult to distinguish between human or AI authored text:

​

Robots are already generating “human-free stories.” The Washington Post, Associated Press, and LA Times use in-house software to create news and social media posts. A growing number of commercial products automate the production of industry reports, email, and marketing copy, and more, and provide options to “humanize” style and tone (Hong).

​

This shows that a person is not always behind the screen, posting for a company. Instead, AI generates content using analytics to receive positive engagement. If AI can write posts or even a novel to the same quality of real person, what reasoning is there to keep paying someone when a robot can do it better?

​

The copyeditor begins their task after a development editor has already fulfilled their role with the manuscript. This means that as such, they can absorb the task of checking for nuance of language, the one thing AI is not capable of, yet. Complex content is not ready to be dealt with using NLP (natural language processing), and we are yet to bring a workable real-world situation live without human intervention; NLP with assisted copyediting is much more practical and reliable (Tan). NLP is a form of AI that is the computer’s ability to study human language both as it is written and spoken, aka linguistics. Just as in Nizinsky’s example, Tan suggests that this AI be implemented by publishers, with minimal human copyediting work done over.