What real writers say about using (or not using) AI

Most of what you read about writers using AI has a bias. It’s filled with writers who like to shout about what they think. Social media algorithms reward the shouting, and you get one of two impressions, depending on your friends: Everyone is using AI for everything, or AI is going to cause the downfall of civilization.
What real writers say
My writer survey, sponsored by Gotham Ghostwriters, is still going on. We already have more than 1,000 completed surveys. If you haven’t taken 15 minutes to tell us what you think, please take a moment to do that now.
I’d love to share the data with you, but like any good researcher I’ll wait until it’s complete.
But what I can tell you is that in a survey that features so many rank-and-file writers — technical writers, marketing writers, authors, speechwriters, and more — we’ve got an incredible diversity of experience and opinion. People’s attitudes are all over the map.
And I can share some of the verbatim comments from open-ended fields in the survey.
On the one hand, we heard comments about how AI users are finding it extremely helpful and useful — not for writing finished text, but for becoming far more productive. These are all real comments volunteered by writers who use AI in their work:
I started writing again after decades. It also makes some things a lot faster.
I think it will help us grow by giving us alternative sources of revenue, like AI generated audiobooks. Plus, it will lessen the time I spend overall, increasing my per hour rate.
AI helps my research faster, cutting down at least 50% of my research time in many cases.
AI gives me writing options. For general tasks such as socials, back of brochure blurbs, etc., it helps to have the “quick and dirty” writing available. I feed it my experts, initial copy, and related info. The result is fast and usable. Still, I edit most AI-generated copy. It’s not plug and play here.
[AI helps me generate] better quality text more quickly, leaving more time for other professional and non-work related activities. Taking the “grind” out of research. Fitting work into the hours I’m actually getting paid for.
But for every AI proponent, there is an AI opponent. And these writers who object to AI were quite eloquent about it.
The rise of AI has shown me what people in general really think of writing — they hate doing it, and they don’t see it as valuable work. They also lack the skills to see when writing is good vs when it is garbage.
AI often gets answers wrong, so it isn’t useful for copy editing at all. In writing, AI is devoid of creativity and the human touch that makes work meaningful. There is no place for AI in the creative writing sphere, in my opinion. I also don’t want to use a tool that has such a vast negative environmental impact, as well as one that is morally wrong; stealing from so many writers without permission to train the tool is bad enough for me to never touch it.
In whatever writing I’m paid to do, I find that the discovery and revision process, coming from me directly, is the heart of any project. Using AI, with its promise of greater productivity and efficiency (neither of which I need), is like having too many cooks in the kitchen or writing by committee. No thanks.
I don’t like them, I don’t find them helpful, they aren’t reliable or accurate. They are built on plagiarism of the hard work done by me and other writers like me and I don’t want anything to do with that system. But most of all…why on earth would I want to outsource my creativity, my critical thinking, and my ideas out to an app owned by a company? I love writing. I want to do the thing I love with my own hands and my own brain.
You mean the sociopathic plagiarism machine that’s prone to flattery, racism, AND delusions, weakens critical thinking, and is an environmental disaster? Hard pass.
I believe it will damage it greatly, and if strong steps aren’t taken very soon it may be irreparable. I know that there are things under the hood that are probably using LLM knowledge or generated content but I’m doing my best to use every tool in my technical arsenal to make certain that all my output avoids any use of AI.
AI will destroy unique writing voice. By its very nature, AI is moderating: it distills unique voice together into something palatable and more easily-digestible. Companies and authors who utilize AI will lose the unique traits of their brand voice or narrative voice, respectively. Everything will begin to sound the same.
A fascinating example of the pros and cons of using AI
Here’s what I did to find these verbatims. I exported all the verbatim comments from SurveyMonkey into a spreadsheet and asked my assistant on this project to find the most interesting comments, both pro-AI and anti-AI. Since there were well over 1,000 individual comments, I encouraged her to use AI to find the interesting ones.
She used this entirely prompt with ChatGPT to find the most interesting verbatims:
Here is an Excel workbook containing direct quotes (“verbatims”) from professional writers answering questions from a survey about their opinion of using artificial intelligence tools for professional writing. Each worksheet is for one question (the text of which is written in cell A1) with the respondents’ answers to that question in column C starting at row 5. Extract 14 of the most provocative and reactionary verbatims from the workbook, seven which have a positive outlook on AI use for writing and seven which have a negative outlook on AI for writing. Provide the verbatims unedited in a table with a column for the worksheet and a column for the cell number so I can verify their accuracy.
ChatGPT did a very nice job of finding interesting quotes and identifying the cell number they were in. But despite the word “unedited,” it changed the quotes. My assistant asked for the cell numbers specifically so we could verify the quotes, and sure enough, its version of the quotes was more bland and sanitized than the actual quotes. Reminding it of the directive to gather unedited quotes made no difference — it still paraphrased them.
This, in a nutshell, is the tradeoff with AI: it’s much faster than doing tedious work by hand, but you have to check its work for accuracy.
I’d love to hear your opinion. Please take the survey if you haven’t already.
Josh, I’m a writer and who loves and benefits from your blog. You might find an AI-related article I recently wrote interesting, It offers a different “take” on how humans (including writers 😍) might fundamentally reorient their online conduct and experience with chatbots. https://www.lionsroar.com/what-if-buddhists-engaged-ai-as-part-of-practice/
Keep up the great work! It matters and is much appreciated! 🙏🏻
AI is a tool. That’s it. No more, no less. There is a propensity (perhaps more accurately called an “addiction”) in this country (if not the world) for us collectively to become obsessed with the latest shiny object. Along comes AI and we see it as a hammer; we automatically think, “This is the perfect screwdriver to solve all my problems!” Yes, use AI appropriately, and it will be a helpful tool. Use it inappropriately, and chaos will ensue.
I’m currently disgusted with an international software manufacturer, whose products I’ve used (while an employee) quite happily for over 30 years. Now, when as a freelancer I bought a year’s subscription to the program from a certified third-party vendor, the manufacturer closed my license after a few weeks. It blames me for not certifying the third-party vendor and not supplying them with my license for the subscription (which they – the manufacturer – never provided to me). I’m not sure how I am responsible for these things – my responsibility was merely to purchase the subscription and use it to provide services for my clients.
After wrangling with the manufacturer’s customer support department for three weeks, I realized that I was arguing with AI. The repetition of content in the emails, the repeated requests for me to provide the manufacturer with materials that I had already provided, the shuttling from one department to another, none of them appropriate – these clued me in.
Employing AI to replace live humans in customer service is NOT a model to implement. It tells me that the manufacturer despises its human customers. This has led to a 40 percent loss in my business income. It burdens my clients, too, because freelancers like me, with industry-specific experience, providing my services to small businesses too small to warrant a full-time employee or thousands-plus for a yearly subscription, are few and far between.