The robots are coming . . . they may even be coming for our jobs! Or, at least, that’s what some recent headlines may have people believing about artificial intelligence (AI) software like ChatGPT, or AI proofreading apps, such as Proofcheck. But is it true?
“I’m a copywriter. I’m pretty sure artificial intelligence is going to take my job,” writes Henry Williams for The Guardian in an Op-Ed earlier this year. The author writes, “My amusement turned to horror: it took ChatGPT 30 seconds to create, for free, an article that would take me hours to write.” Alan Henry, author of Seen, Heard, and Paid: The New Work Rules for the Marginalized (Penguin Random House, 2022) and Special Projects Editor at Wired, spoke with the New York Book Forum for this white paper, and he counters the idea that AI will replace editors, copyeditors, and proofreaders in the near future.
Fear is contagious, though, and there is undoubtedly fear in the water when it comes to newly emerging software like AI-powered writing and editing software. But could some of this fear be unfounded? Yes. Is it really all doom and gloom? No. One needs only look at declining e-book sales industry-wide—a format which was originally believed to be the beginning of the end of physical books—to understand that fears like these can be unfounded and overblown.
Historically, humans tend to fear change. (A 2016 article in The Washington Post summarized this sentiment in its headline: “Humans once opposed coffee and refrigeration. Here’s why we often hate new stuff.”) And an 1895 interview with Thomas A. Edison, inventor of the incandescent light bulb and phonograph, further illuminates this, demonstrating how something that is believed to be growing obsolete (a horse, a bicycle), is actually safe and sound. Edison reflected on “horseless carriages” and what their invention meant for society. Those horseless carriages were automobiles. In an interview reported by The Saturday Evening Post, the inventor says:
“Talking of horseless carriage suggests to my mind that the horse is doomed. The bicycle, which, 10 years ago, was a curiosity, is now a necessity. It is found everywhere. Ten years from now you will be able to buy a horseless vehicle for what you would pay today for a wagon and a pair of horses. The money spent in the keep of the horses will be saved and the danger to life will be much reduced.”
If the horse was “doomed” due to cars, are editors doomed because of AI?
Simply put: No.
What Does AI like ChatGPT Do?
Let’s start at the beginning. What does this AI software do, who created it, and why? ChatGPT is the name of an AI chatbot that launched in November 2022 and immediately began causing quite a stir. Its incredible ability to produce original pieces of writing with a conversational tone, using articles and content pulled from online, astounded many. And it was also worrying to teachers, editors, and other professionals. However, it was noted by users that ChatGPT had uneven factual accuracy. It also doesn’t have the ability to cite its sources.
ChatGPT was developed by OpenAI, a U.S.-based AI research lab headquartered in San Francisco. OpenAI’s founders include Elon Musk, Sam Altman, and others. Microsoft is among its corporate investors. ChatGPT is currently valued at $29 billion.
Chatbots like ChatGPT are designed to imitate human conversation. But ChatGPT does more than just that: it can write short stories, compose poetry and song lyrics. And concerningly for teachers: it can even answer test questions and write student essays.
The Limitations & Biases of AI
While AI like ChatGPT can do a lot, the limitations are well-documented, too. Along with factual inaccuracies, it also produces incorrect, nonsensical writing for some users. The AI software also has biases around race and sex, which remain problematic. AI is imperfect; humans created it and thus it contains human error and humanly flawed thinking.
It’s also very noteworthy that over 95% of the time, when used, ChatGPT is detectable. ChatGPT only speaks one language: English. Though it was recently announced that Icelandic would be the second language the software will learn how to speak. AI proofreading app Proofcheck, on the other hand, is multilingual, offering proofreading support in dozens of languages.
(A 2022 Publishers Lunch article reported that Proofcheck is receiving venture capital funding and continues to generate interest from investors and users. The app was even tested in production departments at the Big Five publishers in the U.S.)
AI-generated image created using the prompt “a robot reading a book”
AI as a Tool, Not a Threat
Wired editor and author of Seen, Heard, and Paid, Alan Henry tells New York Book Forum, “In its current form, AI can be a powerful tool, but right now it’s still that: a tool, that needs to be guided, observed, and managed by human hands.”
In an interview with the Nieman Journalism Lab, the first-ever AI editor at the Financial Times, Madhumita Murgia, recognizes the global shift around AI and how it might be used in the future. Murgia echoes Henry’s thoughts about AI being merely a tool. She says, “I want to make AI more understandable and accessible to our readers, so it doesn’t feel like magic but merely a tool that they can wield.”
Henry points out how AI is fallible and has biases. “AI text generation systems are still prone to simple elementary errors, inherent biases due to poorly engineered datasets, and hallucinations where the tool isn’t aware of decisions it made or processes it executed only moments before.”
The New York-based journalist and author emphatically says AI will not replace editors, copyeditors, or proofreaders anytime soon. An article published by Inc. magazine argues that idea, saying, “human-led AI is the future of content [writing].” But Henry’s assessment notes how flawed this tool is and how it won’t replace humans as a result. He explains, “So while I have hope that eventually it could be used by skilled professionals to make their work easier and more thorough, it absolutely won’t replace human editors, copyeditors, or proofreaders in the near future.”
While he’s not concerned about editorial departments being replaced by AI software, he does have some other concerns. Henry explains, “I am however concerned that in the AI industry gold rush we’re in the middle of, these tools will be used to replace skilled professionals with tools that are nowhere near their competency. I can only hope that as the technology evolves, it’ll serve to enhance the way we work instead of uprooting even more skilled workers from roles where they’re critically necessary.”
Existing AI Editing Software
Many writers and journalists use some form of editing software to help them elevate their already-good copy and catch grammar mistakes, too . . . before their editor does.
Grammarly, for instance, is an online writing assistant, and it has been around since 2009. Its AI-powered suggestions help people looking to strengthen their writing and editorial skills. AI-powered editing and writing tools like Grammarly have existed for over ten years yet haven’t had a substantive negative impact thus far.
Journalisttoolbox.org has an entire page dedicated to AI tools for writers and editors. The organization lists the different software on the market and what each one does. For instance, there’s Jasper (an AI writing tool), QuillBot (an AI writing tool that paraphrases), Craft.do (document-creating with AI features), WordTuneSpices (an advanced word processing tool used by editors), and DeepL Write (which improves copy with its AI).
ChatGPT caused a big stir because of its capabilities, but frankly, similar AI-powered products were already on the market, and there’s been no recorded job loss because of them.
Concerns about AI’s Capabilities
While AI-powered software like Grammarly and ChatGPT haven’t put huge swaths of editors and writers out of jobs yet, there are prominent people who do caution about the power and potential harmful impact of AI.
Renowned theoretical physicist Stephen Hawking is one such voice. Hawking told the BBC in an earlier interview, “The development of full artificial intelligence could spell the end of the human race.” The scientist warns how AI “would take off on its own, and re-design itself at an ever-increasing rate.” Hawking notes that humans have a much slower biological evolutionary process, unlike AI and robots, who can learn and evolve far more rapidly.
While it remains to be seen if AI will pose any real threat, one thing is—for now—certain: AI isn’t going anywhere, and neither is old-fashion, human-led editing, copyediting, and proofreading.
Books always have (and always will) require a human touch: it’s what draws us readers to them.